PASS Summit West (Seattle)
All Sessions
-
Pre-Conference
Building an AI-Enabled Fabric Practice
Ginger Grant
Want to move beyond theory and actually embed AI and automation into your Microsoft Fabric workflows? This session shows how to turn AI’s promised productivity gains into disciplined, repeatable practices you can implement immediately in Microsoft Fabric. We will walk through how to set up an AI-ready environment on your own laptop, along with the tools, libraries, and patterns needed so you can follow along and reuse everything yourself. The session covers how to build prompt libraries, reusable patterns, and guardrails so your results are consistent instead of hit-or-miss. We will also use Fabric APIs and CLIs to turn those prompts into agentic processes so you can implement strategies beyond cutting and pasting from chat windows. To ensure governance and good CI/CD practices are included, the session shows how to automate quality checks with GitHub Actions and enforce standards using AI-assisted code generation and review. These processes can ensure hallucinations don’t see production and ensure improve that you can catch issues before your users do. Using real examples, we will walk through how to move from written specifications to fully implemented models, moving from inception to deployment pipelines. Attendees will leave with a structured framework, practical automation techniques, and repeatable methods to improve quality, consistency, and delivery speed across your Fabric environment.
Pre-Conference
Building an AI-Enabled Fabric Practice
Ginger Grant
Desert Isle Group
Session Goals: • Learn the tools, techniques and setup requirements to needed to create an AI environment for Fabric. • Understand the process to build repeatable AI processes to implement Fabric requirements. • Recognize how to implement testing and CI/CD elements to improve the quality of your Fabric output.
Session Prerequisites: • Some experience with Fabric is required.
Track: Analytics
Level: Level 200
Theme: AI + Data
-
Monday -
9:00 AM-4:30 PM -
Room: 301 Ashnola
Want to move beyond theory and actually embed AI and automation into your Microsoft Fabric workflows? This session shows how to turn AI’s promised productivity gains into disciplined, repeatable practices you can implement immediately in Microsoft Fabric. We will walk through how to set up an AI-ready environment on your own laptop, along with the tools, libraries, and patterns needed so you can follow along and reuse everything yourself. The session covers how to build prompt libraries, reusable patterns, and guardrails so your results are consistent instead of hit-or-miss. We will also use Fabric APIs and CLIs to turn those prompts into agentic processes so you can implement strategies beyond cutting and pasting from chat windows. To ensure governance and good CI/CD practices are included, the session shows how to automate quality checks with GitHub Actions and enforce standards using AI-assisted code generation and review. These processes can ensure hallucinations don’t see production and ensure improve that you can catch issues before your users do. Using real examples, we will walk through how to move from written specifications to fully implemented models, moving from inception to deployment pipelines. Attendees will leave with a structured framework, practical automation techniques, and repeatable methods to improve quality, consistency, and delivery speed across your Fabric environment. -
-
Pre-Conference
Dev-Prod Demon Hunters: Finding the Real Cause of Production Slowness
Brent Ozar
Production is slow. Development is fast. The same query runs in both. Somewhere between the two, a performance demon is hiding—and this session is about hunting it down. Inspired by Brent Ozar's love of the K-Pop Demon Hunters theme song, this class is delivered almost entirely as live demos, not slides. Brent Ozar will run real queries against two environments labeled “dev” and “prod,” then work through them exactly the way an experienced DBA would in the real world: comparing server settings, analyzing execution plans, and uncovering the subtle differences that led SQL Server to make different decisions. Each “hunt” reveals another demon—statistics, configuration, data distribution, or plan choice—and shows how easily a test environment can lie. Along the way, Brent will demonstrate practical techniques you can use immediately: running sp_Blitz to surface meaningful environment differences, comparing execution plans to understand *why* SQL Server behaved differently, and making targeted changes to development so it better reflects production reality. By the end, you’ll understand how to stop guessing, stop blaming the engine, and follow the clues that lead to the truth—because when dev and prod finally move in sync, that’s when performance goes golden.
Pre-Conference
Dev-Prod Demon Hunters: Finding the Real Cause of Production Slowness
Brent Ozar
Brent Ozar Unlimited
Session Goals: • Discover what caused query plans to vary from production. • Learn how to quickly assess environment differences that would cause query plan changes. • Understand how to change dev to more closely match prod.
Session Prerequisites: • You should already be comfortable writing queries, reading execution plans, and using the First Responder Kit to gather data about your server's wait stats and health.
Track: Database Management
Level: Level 300
Theme: Cloud + Multi-DB
-
Monday
-
9:00 AM-4:30 PM
-
Room: 402 Chiliwack
Production is slow. Development is fast. The same query runs in both. Somewhere between the two, a performance demon is hiding—and this session is about hunting it down. Inspired by Brent Ozar's love of the K-Pop Demon Hunters theme song, this class is delivered almost entirely as live demos, not slides. Brent Ozar will run real queries against two environments labeled “dev” and “prod,” then work through them exactly the way an experienced DBA would in the real world: comparing server settings, analyzing execution plans, and uncovering the subtle differences that led SQL Server to make different decisions. Each “hunt” reveals another demon—statistics, configuration, data distribution, or plan choice—and shows how easily a test environment can lie. Along the way, Brent will demonstrate practical techniques you can use immediately: running sp_Blitz to surface meaningful environment differences, comparing execution plans to understand *why* SQL Server behaved differently, and making targeted changes to development so it better reflects production reality. By the end, you’ll understand how to stop guessing, stop blaming the engine, and follow the clues that lead to the truth—because when dev and prod finally move in sync, that’s when performance goes golden. -
-
Pre-Conference
Adding PostgreSQL to Your SQL Server Skill Set
Grant Fritchey
Pat Wright
More organizations are adding PostgreSQL to their technology stack than ever before. The challenge is that they are not immediately replacing their existing technology, which means more and more people need to understand both SQL Server and PostgreSQL. This session is explicitly designed to support people who already know SQL Server in their journey to add PostgreSQL to their skill set. The session covers the areas of overlap between the two platforms, as well as all the differences that can make learning PostgreSQL a challenge. Not only does this all-day session teach PostgreSQL, but it also explores tooling, documentation, the cloud, and other resources to help on the journey of adding PostgreSQL to an existing SQL Server skill set.
Pre-Conference
Adding PostgreSQL to Your SQL Server Skill Set
Grant Fritchey
Redgate
Pat Wright
Redgate
Session Goals: • Learn about the differences in language between SQL Server and PostgreSQL. • Understand what's different and unique in how PostgreSQL operates.
Session Prerequisites: • General understanding of SQL Server in order to understand the mapping.
Track: Database Management
Level: Level 200
Theme: Cloud + Multi-DB
-
Monday
-
9:00 AM-4:30 PM
-
Room: 307 Methow
More organizations are adding PostgreSQL to their technology stack than ever before. The challenge is that they are not immediately replacing their existing technology, which means more and more people need to understand both SQL Server and PostgreSQL. This session is explicitly designed to support people who already know SQL Server in their journey to add PostgreSQL to their skill set. The session covers the areas of overlap between the two platforms, as well as all the differences that can make learning PostgreSQL a challenge. Not only does this all-day session teach PostgreSQL, but it also explores tooling, documentation, the cloud, and other resources to help on the journey of adding PostgreSQL to an existing SQL Server skill set. -
-
Pre-Conference
T-SQL That Doesn't Suck: Solving Performance and Concurrency Problems
Erik Darling
You know how to write T-SQL that works. It compiles, it returns rows, nobody's filed an incident yet. But "works" is doing a lot of heavy lifting in that sentence, and production has a way of finding out what you got away with in dev. In this full-day, demo-packed session, Erik Darling walks you through the T-SQL that silently bleeds your servers dry and shows you how to fix it without rewriting the whole app. You'll tackle the performance killers that don't show up until scale does: implicit conversions, non-sargable predicates, parameter sniffing traps, and joins that look fine in an execution plan until they absolutely aren't. You'll learn when temp tables beat table variables, when CTEs help versus when they just make you feel organized, and how to write window functions that don't spill to tempdb. But performance is only half the fight. We'll dig into concurrency problems that turn a Tuesday into a war room: blocking chains, isolation level surprises, and data modifications that hold locks like they're paying rent. You'll learn DML patterns that play well with others. Along the way, we'll put AI-generated T-SQL on the table. Not to dunk on it, but because it's showing up in pull requests and you need to know where it falls apart and where it saves you time. If you've ever inherited a stored procedure that made you question your career choices, this is your day. Come write T-SQL you'd be proud to put your name on.
Pre-Conference
T-SQL That Doesn't Suck: Solving Performance and Concurrency Problems
Erik Darling
Darling Data
Session Goals: • Recognize and fix performance anti-patterns across joins, predicates, temp objects, CTEs, and window functions. • Understand and resolve concurrency issues including blocking, isolation levels, and lock-friendly DML. • Evaluate AI-generated T-SQL and learn to use AI tools without losing your edge.
Session Prerequisites: • You should be comfortable writing T-SQL: SELECT, JOIN, WHERE, GROUP BY, and basic stored procedures. • If you can read an execution plan without panicking, you're ready. • This session picks up where the fundamentals leave off.
Track: Development
Level: Level 300
Theme: AI + Data
-
Monday
-
9:00 AM-4:30 PM
-
Room: 302 Beckler
You know how to write T-SQL that works. It compiles, it returns rows, nobody's filed an incident yet. But "works" is doing a lot of heavy lifting in that sentence, and production has a way of finding out what you got away with in dev. In this full-day, demo-packed session, Erik Darling walks you through the T-SQL that silently bleeds your servers dry and shows you how to fix it without rewriting the whole app. You'll tackle the performance killers that don't show up until scale does: implicit conversions, non-sargable predicates, parameter sniffing traps, and joins that look fine in an execution plan until they absolutely aren't. You'll learn when temp tables beat table variables, when CTEs help versus when they just make you feel organized, and how to write window functions that don't spill to tempdb. But performance is only half the fight. We'll dig into concurrency problems that turn a Tuesday into a war room: blocking chains, isolation level surprises, and data modifications that hold locks like they're paying rent. You'll learn DML patterns that play well with others. Along the way, we'll put AI-generated T-SQL on the table. Not to dunk on it, but because it's showing up in pull requests and you need to know where it falls apart and where it saves you time. If you've ever inherited a stored procedure that made you question your career choices, this is your day. Come write T-SQL you'd be proud to put your name on. -
-
Pre-Conference
Performance Troubleshooting Using Waits and Latches
Paul Randal
SQL Server performance is off – what do you do? One of the first things you should check when investigating performance issues are wait and latch statistics, as these often point you in the right direction for further analysis. Unfortunately many people misinterpret what SQL Server is telling them and jump to conclusions about how to solve the problem ('knee-jerk performance tuning'). In this full-day workshop, you'll learn how to investigate and interpret wait and latch statistics – easy, practical techniques you can take home and start using immediately. You'll also learn what a myriad of wait and latch types actually mean and how you can investigate lightly-documented types to gain insight into what causes them. Don't waste time when it comes to performance troubleshooting; wait statistics are the fastest route to understanding your problems and this workshop will help you get there faster.
Pre-Conference
Performance Troubleshooting Using Waits and Latches
Paul Randal
SQLskills
Session Goals: • Clear understanding of what waits and latches are and how they can guide performance tuning. • Ability to run waits-related DMV scripts and know how to proceed with the results. • Understanding of common wait and latch types, what they mean, and how to react to them.
Session Prerequisites: • Familiarity with terms and concepts like transactions, locks, indexes, DMVs. Definitely not a 100-level, complete beginner workshop.
Track: Database Management
Level: Level 300
Theme: AI + Data
-
Monday
-
9:00 AM-4:30 PM
-
Room: 401 Chelan
SQL Server performance is off – what do you do? One of the first things you should check when investigating performance issues are wait and latch statistics, as these often point you in the right direction for further analysis. Unfortunately many people misinterpret what SQL Server is telling them and jump to conclusions about how to solve the problem ('knee-jerk performance tuning'). In this full-day workshop, you'll learn how to investigate and interpret wait and latch statistics – easy, practical techniques you can take home and start using immediately. You'll also learn what a myriad of wait and latch types actually mean and how you can investigate lightly-documented types to gain insight into what causes them. Don't waste time when it comes to performance troubleshooting; wait statistics are the fastest route to understanding your problems and this workshop will help you get there faster. -
-
Pre-Conference
Learning the Cloud: From Managing Resources to Designing Solutions
Joey D'Antoni
With over 200 services in Azure, trying to learn the platform can feel overwhelming. You've probably built a virtual machine or deployed an Azure SQL Database, but have you thought about your network design, tagging strategies, or how AI services fit into your architecture? In this all-day session, we'll go well beyond basic DBA and sysadmin tasks and into the mindset of an architect. Topics covered will include: Azure Networking Azure and Entra ID Security How to automate processes in Azure A deep dive on storage performance and costs Azure AI services including Microsoft Foundry, AI Search, and Cognitive Services We'll start with Azure's foundational infrastructure and how to manage availability. From there, we'll cover how security, authentication, and authorization work with Entra ID and how to implement them in your applications. You'll learn how to choose the right data and AI services for your workloads, how to secure them, and when to reach for tools like Azure OpenAI Service or AI Search to add intelligence to your solutions. We'll also dig into supporting services like Azure Key Vault, Azure Functions, Logic Apps, and Automation that help tie everything together. This session will be packed with demos and plenty of open Q&A.
Pre-Conference
Learning the Cloud: From Managing Resources to Designing Solutions
Joey D'Antoni
3Cloud
Session Goals: • Evaluate and design Azure networking, security, and storage architectures beyond basic admin tasks. • Implement authentication and security strategies using Entra across your applications and data • Identify when and how to integrate AI services like OpenAI and AI Search into your solution architecture.
Session Prerequisites: • None.
Track: Architecture
Level: Level 200
Theme: Cloud + Multi-DB
-
Monday
-
9:00 AM-4:30 PM
-
Room: 405 Kachess
With over 200 services in Azure, trying to learn the platform can feel overwhelming. You've probably built a virtual machine or deployed an Azure SQL Database, but have you thought about your network design, tagging strategies, or how AI services fit into your architecture? In this all-day session, we'll go well beyond basic DBA and sysadmin tasks and into the mindset of an architect. Topics covered will include: Azure Networking Azure and Entra ID Security How to automate processes in Azure A deep dive on storage performance and costs Azure AI services including Microsoft Foundry, AI Search, and Cognitive Services We'll start with Azure's foundational infrastructure and how to manage availability. From there, we'll cover how security, authentication, and authorization work with Entra ID and how to implement them in your applications. You'll learn how to choose the right data and AI services for your workloads, how to secure them, and when to reach for tools like Azure OpenAI Service or AI Search to add intelligence to your solutions. We'll also dig into supporting services like Azure Key Vault, Azure Functions, Logic Apps, and Automation that help tie everything together. This session will be packed with demos and plenty of open Q&A. -
-
Lightning Talk
GenAI for Schema Conversion: What It Actually Gets Right (and Wrong)
Marcelo Fernandes
Generative AI is transforming how we approach database schema conversion, promising to automate the tedious, error-prone work that has historically slowed down migration projects. But how well does it actually perform in practice? This lightning session delivers an honest, tool-agnostic assessment of using generative AI for database code conversion. Drawing from real migration projects, we examine specific T-SQL patterns and stored procedures where AI-powered conversion excels — and where it confidently produces code that is subtly wrong. We walk through concrete before-and-after examples: complex joins, cursor-based logic, dynamic SQL, error handling blocks, and data type mappings. For each category, you will see the AI-generated output alongside the manually corrected version, so you can calibrate your own expectations. Beyond individual conversions, we discuss the human review process that remains essential. How do you efficiently validate AI-generated code? What testing strategies catch the failures that look correct at first glance? And how do you build a feedback loop that improves conversion quality over time? This is not a product demo or a marketing pitch. It is a practitioner's perspective on where generative AI genuinely accelerates schema conversion today, where it creates false confidence, and how to build a realistic workflow that leverages AI without blindly trusting it. Set your expectations before your next migration.
Lightning Talk
GenAI for Schema Conversion: What It Actually Gets Right (and Wrong)
Marcelo Fernandes
AWS
Session Goals: • Understand where generative AI reliably accelerates schema conversion and where it introduces subtle errors. • Learn practical strategies for validating and testing AI-generated database code. • Build realistic expectations for incorporating GenAI into your database migration workflow.
Session Prerequisites: • Experience with MSSQL and T-SQL. • Basic understanding of database migration concepts and schema conversion challenges. • Familiarity with stored procedures, functions, and database code objects. • No prior experience with generative AI tools required.
Track: Architecture
Level: Level 100
Theme: AI + Data
-
-
-
Room: TBD
Generative AI is transforming how we approach database schema conversion, promising to automate the tedious, error-prone work that has historically slowed down migration projects. But how well does it actually perform in practice? This lightning session delivers an honest, tool-agnostic assessment of using generative AI for database code conversion. Drawing from real migration projects, we examine specific T-SQL patterns and stored procedures where AI-powered conversion excels — and where it confidently produces code that is subtly wrong. We walk through concrete before-and-after examples: complex joins, cursor-based logic, dynamic SQL, error handling blocks, and data type mappings. For each category, you will see the AI-generated output alongside the manually corrected version, so you can calibrate your own expectations. Beyond individual conversions, we discuss the human review process that remains essential. How do you efficiently validate AI-generated code? What testing strategies catch the failures that look correct at first glance? And how do you build a feedback loop that improves conversion quality over time? This is not a product demo or a marketing pitch. It is a practitioner's perspective on where generative AI genuinely accelerates schema conversion today, where it creates false confidence, and how to build a realistic workflow that leverages AI without blindly trusting it. Set your expectations before your next migration. -
-
Lightning Talk
Native Regex on the Microsoft Data Platform
Edward Pollack
This is a fun and fast-paced run through the native Regex functionality in Azure SQL Database and SQL Server 2025. Gone are the days of complex string-matching and endless LIKE operators! Native Regex has been one of the most requested features in T-SQL for years and this is an opportunity to finally use it to simplify your old code, improve performance, and write more reliable queries!
Lightning Talk
Native Regex on the Microsoft Data Platform
Edward Pollack
Transfinder
Session Goals: • Learn about the many new Regex functions in T-SQL. • Understand how these new functions can simplify code and improve preformance. • Learn how to use these as strict upgrades to existing functions.
Session Prerequisites: • None.
Track: Development
Level: Level 100
-
-
-
Room: TBD
This is a fun and fast-paced run through the native Regex functionality in Azure SQL Database and SQL Server 2025. Gone are the days of complex string-matching and endless LIKE operators! Native Regex has been one of the most requested features in T-SQL for years and this is an opportunity to finally use it to simplify your old code, improve performance, and write more reliable queries! -
-
General Session
The Architecture Gap: Why Your AI Strategy Is Slower Than You Think
Kevin Wilkie
AI isn’t failing because of models, tools, or talent—it’s failing because of architecture. Across enterprises, organizations are investing heavily in AI while quietly running on data platforms that were never designed to support it. The result? Slow delivery, fragile pipelines, rising costs, and “AI initiatives” that never make it to production. In this session, we’ll break down the growing gap between modern AI ambitions and legacy (and even “modernized”) data architectures. You’ll see why lift-and-shift migrations often make the problem worse, how hybrid architectures are emerging as a practical path forward, and what architectural patterns actually enable AI at scale. If your organization is struggling to turn AI strategy into real outcomes, the problem may not be your AI—it may be your platform.
General Session
The Architecture Gap: Why Your AI Strategy Is Slower Than You Think
Kevin Wilkie
Sherpa of Data
Session Goals: • Evaluate if your current data architecture can support AI workloads. • Identify architectural anti-patterns (such as lift-and-shift) that can increase cost and complexity. • Apply hybrid architecture strategies to better align platforms with AI use cases and business outcomes.
Session Prerequisites: • Attendees should have intermediate familiarity with data platforms, including data warehousing and ETL/ELT processes. • Experience with platforms such as SQL Server, Snowflake, and Microsoft Fabric is helpful, but not required.
Track: Architecture
Level: Level 200
Theme: AI + Data
-
-
-
Room: TBD
AI isn’t failing because of models, tools, or talent—it’s failing because of architecture. Across enterprises, organizations are investing heavily in AI while quietly running on data platforms that were never designed to support it. The result? Slow delivery, fragile pipelines, rising costs, and “AI initiatives” that never make it to production. In this session, we’ll break down the growing gap between modern AI ambitions and legacy (and even “modernized”) data architectures. You’ll see why lift-and-shift migrations often make the problem worse, how hybrid architectures are emerging as a practical path forward, and what architectural patterns actually enable AI at scale. If your organization is struggling to turn AI strategy into real outcomes, the problem may not be your AI—it may be your platform. -
-
General Session
Modernize Without Moving (Yet): SQL Migration and Management with Azure Arc
Ben Weissman
Modernizing your SQL Server estate doesn’t have to start with a full migration to the cloud. With Azure Arc, you can bring cloud-based management, governance, and innovation to your existing environments, whether they run on-premises, at the edge, or in other clouds. In this session, we explore how Azure Arc for SQL Server enables centralized visibility, policy-based governance, and streamlined operations across your entire data estate. You’ll learn how to onboard and manage SQL Server instances at scale, assess your environment for migration readiness, and leverage Arc-enabled services to reduce friction in your modernization journey. We’ll also connect the dots between management and migration: how Arc helps you identify the right candidates, prepare workloads, and incrementally move toward Azure on your terms and timeline. Expect practical demos, real-world scenarios, and clear guidance on how to modernize step by step, without disrupting what already works. Whether you’re planning a migration or just want better control today, Azure Arc provides the bridge between where you are and where you want to be.
General Session
Modernize Without Moving (Yet): SQL Migration and Management with Azure Arc
Ben Weissman
Solisyon
Session Goals: • Understand how to enhance on-prem SQL Server functionality through Arc. • Build a central management pane for your entire SQL estate, no matter where it is. • Facilitate easy migrations.
Session Prerequisites: • Basic understanding of SQL Server and Azure terminology.
Track: Architecture
Level: Level 200
Theme: Cloud + Multi-DB
-
-
-
Room: TBD
Modernizing your SQL Server estate doesn’t have to start with a full migration to the cloud. With Azure Arc, you can bring cloud-based management, governance, and innovation to your existing environments, whether they run on-premises, at the edge, or in other clouds. In this session, we explore how Azure Arc for SQL Server enables centralized visibility, policy-based governance, and streamlined operations across your entire data estate. You’ll learn how to onboard and manage SQL Server instances at scale, assess your environment for migration readiness, and leverage Arc-enabled services to reduce friction in your modernization journey. We’ll also connect the dots between management and migration: how Arc helps you identify the right candidates, prepare workloads, and incrementally move toward Azure on your terms and timeline. Expect practical demos, real-world scenarios, and clear guidance on how to modernize step by step, without disrupting what already works. Whether you’re planning a migration or just want better control today, Azure Arc provides the bridge between where you are and where you want to be. -
-
General Session
Git Good: GitHub and GitHub Actions for the Rest of Us
Joey D'Antoni
LLMs are changing how we write code, manage infrastructure, and ship software, making source code management and DevOps skills more valuable than ever. If DevOps has always seemed confusing or mysterious, you're not alone. In this session, you will: – Learn Git and GitHub basics for storing and managing your code – Get comfortable collaborating in repositories – Use GitHub Actions to automate code, cloud, and database deployments – How GitHub Actions are basically just shell scripts – Understand how to ship software faster with fewer problems You'll leave ready to automate, collaborate, and deliver your projects with confidence.
General Session
Git Good: GitHub and GitHub Actions for the Rest of Us
Joey D'Antoni
3Cloud
Session Goals: • Understand how to use GitHub for source code management, including repositories, branches, and commits. • Build and configure GitHub Actions workflows to automate code, cloud, and database deployments. • Apply CI/CD best practices to improve build quality and ship software more reliably.
Session Prerequisites: • None.
Track: Development
Level: Level 200
Theme: AI + Data
-
-
-
Room: TBD
LLMs are changing how we write code, manage infrastructure, and ship software, making source code management and DevOps skills more valuable than ever. If DevOps has always seemed confusing or mysterious, you're not alone. In this session, you will: – Learn Git and GitHub basics for storing and managing your code – Get comfortable collaborating in repositories – Use GitHub Actions to automate code, cloud, and database deployments – How GitHub Actions are basically just shell scripts – Understand how to ship software faster with fewer problems You'll leave ready to automate, collaborate, and deliver your projects with confidence. -
-
Learning Pathway
Adapting for AI Readiness as a Data Professional | Keeping Up Without Falling Down: AI's Future for Data Professionals
Matt Gordon
The first two sessions in this learning pathway cover essential AI-related knowledge for data professionals and then how to apply that knowledge to your work. The final session in the learning pathway covers what we do next and where we go from here. Everyone who works in technology is feeling the pressure to keep up with every new development in AI – and most of us feel like we're failing at that effort, or at least falling behind. This session will focus on where we think AI will take us – and our job responsibilities – in the future. Leaning on the access that Microsoft MVPs have to what Microsoft is thinking about the future of their data platform offerings, we have put together a guide for what we think we should learn, what we think we should do with that newfound knowledge, and how to take care of ourselves in an increasingly turbulent data professional landscape. Our jobs aren't going away, but we need to make sure we're equipped for where they are headed. Join us to help start on that journey.
Learning Pathway
Adapting for AI Readiness as a Data Professional | Keeping Up Without Falling Down: AI's Future for Data Professionals
Matt Gordon
Lean TECHniques
Session Goals: • Learn what to learn and what skills to develop to keep up with the AI-driven evolution of our jobs as data professionals. • Learn how to apply that knowledge and those newfound skills to your work now. • Learn how to take care of ourselves in an increasingly turbulent technology landscape.
Session Prerequisites: • A basic knowledge of SQL is preferred.
Track: Analytics
Level: Level 200
Theme: AI + Data
-
-
-
Room: TBD
The first two sessions in this learning pathway cover essential AI-related knowledge for data professionals and then how to apply that knowledge to your work. The final session in the learning pathway covers what we do next and where we go from here. Everyone who works in technology is feeling the pressure to keep up with every new development in AI – and most of us feel like we're failing at that effort, or at least falling behind. This session will focus on where we think AI will take us – and our job responsibilities – in the future. Leaning on the access that Microsoft MVPs have to what Microsoft is thinking about the future of their data platform offerings, we have put together a guide for what we think we should learn, what we think we should do with that newfound knowledge, and how to take care of ourselves in an increasingly turbulent data professional landscape. Our jobs aren't going away, but we need to make sure we're equipped for where they are headed. Join us to help start on that journey. -
-
General Session
Eventhouse: The Hidden Power Behind Fabric Real-Time Intelligence
Frank Geisler
Eventhouse is the analytical engine behind Real-Time Intelligence in Microsoft Fabric—but most implementations barely scratch the surface of what it can do. In this deep technical session, we go beyond the basics and explore how Eventhouse enables high-performance, low-latency analytics on streaming data at scale. You’ll gain insight into how data is ingested, stored, indexed, and queried within Eventhouse, and why its architecture is fundamentally different from traditional data warehouses and lakehouses. We’ll examine how KQL leverages columnar storage, caching, and distributed execution to deliver near-instant query performance—even on high-volume event streams. The session also covers advanced topics such as data retention strategies, partitioning, ingestion optimization, and query tuning. You’ll learn how to design for both performance and cost efficiency, and how to avoid common bottlenecks in real-world deployments. This is a Level 400 session aimed at experienced data professionals who want to understand not just how to use Eventhouse—but how it works under the hood and how to push it to its limits.
General Session
Eventhouse: The Hidden Power Behind Fabric Real-Time Intelligence
Frank Geisler
GDS Business Intelligence GmbH
Session Goals: • Understand the internal architecture of Eventhouse. • Optimize ingestion and query performance for real-time workloads. • Design scalable and cost-efficient Eventhouse solutions.
Session Prerequisites: • Solid Microsoft Fabric experience. • KQL (Kusto Query Language). • Data architecture knowledge.
Track: Analytics
Level: Level 300
Theme: AI + Data
-
-
-
Room: TBD
Eventhouse is the analytical engine behind Real-Time Intelligence in Microsoft Fabric—but most implementations barely scratch the surface of what it can do. In this deep technical session, we go beyond the basics and explore how Eventhouse enables high-performance, low-latency analytics on streaming data at scale. You’ll gain insight into how data is ingested, stored, indexed, and queried within Eventhouse, and why its architecture is fundamentally different from traditional data warehouses and lakehouses. We’ll examine how KQL leverages columnar storage, caching, and distributed execution to deliver near-instant query performance—even on high-volume event streams. The session also covers advanced topics such as data retention strategies, partitioning, ingestion optimization, and query tuning. You’ll learn how to design for both performance and cost efficiency, and how to avoid common bottlenecks in real-world deployments. This is a Level 400 session aimed at experienced data professionals who want to understand not just how to use Eventhouse—but how it works under the hood and how to push it to its limits. -
-
General Session
Azure PowerShell for DBAs: Stop Clicking, Start Scripting
Mikey Bronowski
If you are managing Azure SQL through the portal, you are doing it the slow way. PowerShell gives you the ability to automate the work you repeat every day – provisioning, configuration, backups, monitoring – and get it done in seconds instead of minutes. This session is built around real, usable examples for DBAs who are comfortable with SQL but newer to scripting in Azure.
General Session
Azure PowerShell for DBAs: Stop Clicking, Start Scripting
Mikey Bronowski
DataPopkorn.com
Session Goals: • Automate common Azure SQL management tasks using PowerShell. • Reduce time spent on repetitive portal work with practical, reusable scripts. • Build a foundation for scripting in Azure without needing to be a developer.
Session Prerequisites: • Basic PowerShell and SQL knowledge.
Track: Development
Level: Level 100
Theme: Cloud + Multi-DB
-
-
-
Room: TBD
If you are managing Azure SQL through the portal, you are doing it the slow way. PowerShell gives you the ability to automate the work you repeat every day – provisioning, configuration, backups, monitoring – and get it done in seconds instead of minutes. This session is built around real, usable examples for DBAs who are comfortable with SQL but newer to scripting in Azure. -
-
General Session
Mom/Daughter Duo: DBA Teaches Data Engineer Security Secrets Worth Stealing
Allison Harris
Sharon Reid
Data engineers and database administrators come at systems from very different angles. One is focused on speed and scale. The other is focused on stability, optimization, and protection. When it comes to security, those two worlds can't afford to stay separate. Join this mother/daughter duo as they bring both perspectives together to explore the security lessons data engineers can and should steal from the DBA playbook. Using practical examples and real-world scenarios, they'll walk through the fundamentals that matter most: least privilege access, secrets and credential management, protecting sensitive data in non-production environments, auditing and traceability, and understanding where fast moving pipelines quietly introduce risk. Whether you manage databases, build data pipelines, or lead data initiatives, you'll walk away with concrete strategies for balancing access and performance, and with a clear understanding of why DBAs and data engineers need to be speaking the same security language.
General Session
Mom/Daughter Duo: DBA Teaches Data Engineer Security Secrets Worth Stealing
Allison Harris
NFP
Sharon Reid
NISA Investment Advisors
Session Goals: • Attendees will be able to reimagine security as a process to implement throughout the pipeline rather than a requirement at the end of the project. • Attendees will walk away with practical use cases for enacting security principles in multiple platforms.
Session Prerequisites: • You would need a basic understanding of databases and ETL processes.
Track: Database Management
Level: Level 200
Theme: Security
-
-
-
Room: TBD
Data engineers and database administrators come at systems from very different angles. One is focused on speed and scale. The other is focused on stability, optimization, and protection. When it comes to security, those two worlds can't afford to stay separate. Join this mother/daughter duo as they bring both perspectives together to explore the security lessons data engineers can and should steal from the DBA playbook. Using practical examples and real-world scenarios, they'll walk through the fundamentals that matter most: least privilege access, secrets and credential management, protecting sensitive data in non-production environments, auditing and traceability, and understanding where fast moving pipelines quietly introduce risk. Whether you manage databases, build data pipelines, or lead data initiatives, you'll walk away with concrete strategies for balancing access and performance, and with a clear understanding of why DBAs and data engineers need to be speaking the same security language. -
-
General Session
Leadership Essentials for Data Teams
Aaron Cutshall
Most people think that leadership is about solving problems. However, leadership isn't just about having power, authority, and making the big decisions. It's about taking ownership and responsibility for the consequences of those decisions and being accountable for how you will respond to problems that arise. On top of that, leaders aren’t just judged by what they do, but also what their people do. Leading a data team presents further unique challenges as most data teams consist of people from widely different backgrounds, skills, aptitudes, and attitudes. Plus, todays data teams are often scattered across multiple time zones, geographies, and languages. This session addresses the primary principles of leadership plus tips on how to: • Handle Personality Conflicts • Deal with Incomplete or Missing Requirements • Enforce Standards & Documentation • Foster Growth & Cohesion Leadership isn't just limited to designated leaders either. Team members are also leaders by example, expertise, or by simply being awesome team members! Therefore, if you lead a data team, want to be a leader, or are a member of a data team, this session is for you!
General Session
Leadership Essentials for Data Teams
Aaron Cutshall
SQLRV, LLC
Session Goals: • Understand fundamental leadership principles. • Encourage growth amidst personality conflicts. • Build & enforce standards with incomplete or missing requirements.
Session Prerequisites: • None.
Track: Professional Development
Level: Level 100
-
-
-
Room: TBD
Most people think that leadership is about solving problems. However, leadership isn't just about having power, authority, and making the big decisions. It's about taking ownership and responsibility for the consequences of those decisions and being accountable for how you will respond to problems that arise. On top of that, leaders aren’t just judged by what they do, but also what their people do. Leading a data team presents further unique challenges as most data teams consist of people from widely different backgrounds, skills, aptitudes, and attitudes. Plus, todays data teams are often scattered across multiple time zones, geographies, and languages. This session addresses the primary principles of leadership plus tips on how to: • Handle Personality Conflicts • Deal with Incomplete or Missing Requirements • Enforce Standards & Documentation • Foster Growth & Cohesion Leadership isn't just limited to designated leaders either. Team members are also leaders by example, expertise, or by simply being awesome team members! Therefore, if you lead a data team, want to be a leader, or are a member of a data team, this session is for you! -
-
General Session
Empowering Lakehouse Solutions with Fabric Warehouse
Christian Henrik Reich
Microsoft Fabric Lakehouses receive a great deal of attention, but Fabric Warehouse is often underused or misunderstood in lakehouse architecture. As the platform has evolved, the Warehouse has become a more important part of the broader Fabric story, especially for teams that need strong SQL-based analytics alongside Spark and Lakehouse workloads. This session takes a closer look at Fabric Warehouse, including its engine, performance characteristics, and the role it can play in modern lakehouse solutions. Attendees will learn where the Warehouse fits, how it interoperates with Spark and Lakehouses, and how to think about it as part of a broader architectural design rather than as a separate or competing option. The session also explores how the Warehouse has developed over time and what that means in practice for solution design, performance, and workload placement. The goal is to help attendees understand not only what Fabric Warehouse is, but when and why to use it. Participants will leave with a clearer view of how Fabric Warehouse can strengthen lakehouse solutions and how to apply it more effectively in real-world Fabric architectures.
General Session
Empowering Lakehouse Solutions with Fabric Warehouse
Christian Henrik Reich
Microsoft
Session Goals: • Understand where Fabric Warehouse fits in a Lakehouse architecture. • Learn how it interoperates with Spark and Lakehouses. • Leave with practical guidance on performance, workload placement, and effective use in real Fabric solutions.
Session Prerequisites: • Basic familiarity with data warehousing, SQL, and modern analytics architectures is helpful. • Some knowledge of Microsoft Fabric or Lakehouse concepts will make the session easier to follow. • Deep Fabric Warehouse experience is not required.
Track: Architecture
Level: Level 300
Theme: AI + Data
-
-
-
Room: TBD
Microsoft Fabric Lakehouses receive a great deal of attention, but Fabric Warehouse is often underused or misunderstood in lakehouse architecture. As the platform has evolved, the Warehouse has become a more important part of the broader Fabric story, especially for teams that need strong SQL-based analytics alongside Spark and Lakehouse workloads. This session takes a closer look at Fabric Warehouse, including its engine, performance characteristics, and the role it can play in modern lakehouse solutions. Attendees will learn where the Warehouse fits, how it interoperates with Spark and Lakehouses, and how to think about it as part of a broader architectural design rather than as a separate or competing option. The session also explores how the Warehouse has developed over time and what that means in practice for solution design, performance, and workload placement. The goal is to help attendees understand not only what Fabric Warehouse is, but when and why to use it. Participants will leave with a clearer view of how Fabric Warehouse can strengthen lakehouse solutions and how to apply it more effectively in real-world Fabric architectures. -
-
General Session
Power Hour
Ben Weissman
Power Hour. The session. The myth. The legendary waste of your valuable PASS Summit time. The one session that must be experienced – live and in person – to be believed. If you want to learn critical skills and cutting-edge technologies… please check the schedule for a different session to attend. But if you want to see Power BI, PowerApps, Flow, Power Singalong, and other Power Platform tools used in ways in which their makers never intended, there is place you should rather be. Power Hour delivers wacky, frequently unseemly, and generally funny demos, all with a serious point about the flexibility and power of Microsoft Power Platform services, tools, and technologies. Nearly endless swag, probably. All are welcome. Must not miss. Professional driver on a closed course. Do not attempt without appropriate safety measures in place.
General Session
Power Hour
Ben Weissman
Solisyon
Session Goals: • None
Session Prerequisites: • None
Track: Analytics
Level: Level 200
-
-
-
Room: TBD
Power Hour. The session. The myth. The legendary waste of your valuable PASS Summit time. The one session that must be experienced – live and in person – to be believed. If you want to learn critical skills and cutting-edge technologies… please check the schedule for a different session to attend. But if you want to see Power BI, PowerApps, Flow, Power Singalong, and other Power Platform tools used in ways in which their makers never intended, there is place you should rather be. Power Hour delivers wacky, frequently unseemly, and generally funny demos, all with a serious point about the flexibility and power of Microsoft Power Platform services, tools, and technologies. Nearly endless swag, probably. All are welcome. Must not miss. Professional driver on a closed course. Do not attempt without appropriate safety measures in place. -
-
General Session
Your MSSQL Environment is Vulnerable and I Can Prove It
Fabiano Amorim
SQL Server environments are often assumed to be secure by default. This session challenges that assumption. Starting from a low-privileged SQL login, I will demonstrate how design decisions, legacy features, and overlooked behaviors inside the MSSQL engine can be combined to achieve full instance compromise. The presentation walks through real-world attack paths observed across on-premises and cloud-hosted SQL Server deployments, showing how privilege boundaries can be bypassed, sensitive data exposed, and administrative control obtained without exploiting traditional memory corruption bugs. Beyond the offensive techniques, the focus is on why these issues exist and how they persist in production systems. Attendees will learn how to assess their own environments, recognize high-risk configurations, and apply practical defensive controls to reduce attack surface and prevent privilege escalation before an attacker proves it for them.
General Session
Your MSSQL Environment is Vulnerable and I Can Prove It
Fabiano Amorim
Pythian
Session Goals: • Understand how MSSQL privilege escalation works across on-prem and cloud environments. • Identify default features and configurations that introduce security risk. • Apply practical techniques to detect, reduce attack surface, and prevent escalation.
Session Prerequisites: • Basic understanding of MSSQL is recommended. • Familiarity with concepts such as DB roles, permissions, triggers, jobs will help attendees follow the attack paths and defensive strategies, though key concepts will be explained during the session.
Track: Database Management
Level: Level 200
Theme: Security
-
-
-
Room: TBD
SQL Server environments are often assumed to be secure by default. This session challenges that assumption. Starting from a low-privileged SQL login, I will demonstrate how design decisions, legacy features, and overlooked behaviors inside the MSSQL engine can be combined to achieve full instance compromise. The presentation walks through real-world attack paths observed across on-premises and cloud-hosted SQL Server deployments, showing how privilege boundaries can be bypassed, sensitive data exposed, and administrative control obtained without exploiting traditional memory corruption bugs. Beyond the offensive techniques, the focus is on why these issues exist and how they persist in production systems. Attendees will learn how to assess their own environments, recognize high-risk configurations, and apply practical defensive controls to reduce attack surface and prevent privilege escalation before an attacker proves it for them. -
-
General Session
Always On Availability Groups and Failover Clustering Internals
Edwin M Sarmiento
Windows Server Failover Clustering (WSFC) is the platform that makes SQL Server Always On Availability Groups highly available. There are external interactions between WSFC and SQL Server that you as a DBA cannot monitor from within SQL Server. You need to know the different components that make up the WSFC and how they work to properly deploy and manage Always On Availability Groups for high availability. In this session, you will dive into the internals of the WSFC, the SQL Server cluster resource DLL, and how they work together to make your SQL Server databases highly available. You will analyze the cluster error log for identifying the root cause of an unexpected or planned failover. Finally, you will learn administrative tasks that you're not supposed to do outside of SQL Server.
General Session
Always On Availability Groups and Failover Clustering Internals
Edwin M Sarmiento
15C
Session Goals: • Learn the interaction between SQL Server and the Windows Server Failover Cluster to achieve high availability. • Avoid performing administrative tasks that jeopardizes availability of SQL Server databases. • How to use the cluster error log with SQL Server logs to troubleshoot availability issues.
Session Prerequisites: • Attendees should have experience deploying SQL Server Always On Availability Groups on Windows Server Failover Clusters.
Track: Database Management
Level: Level 400
-
-
-
Room: TBD
Windows Server Failover Clustering (WSFC) is the platform that makes SQL Server Always On Availability Groups highly available. There are external interactions between WSFC and SQL Server that you as a DBA cannot monitor from within SQL Server. You need to know the different components that make up the WSFC and how they work to properly deploy and manage Always On Availability Groups for high availability. In this session, you will dive into the internals of the WSFC, the SQL Server cluster resource DLL, and how they work together to make your SQL Server databases highly available. You will analyze the cluster error log for identifying the root cause of an unexpected or planned failover. Finally, you will learn administrative tasks that you're not supposed to do outside of SQL Server. -
-
General Session
The Recovering Shy Person's Guide to Putting Yourself Forward
Lenore Flower
Have you ever found yourself holding back from tackling a new project, submitting that conference proposal, or publishing that blog post because the "what ifs" just became too much? Everyone holds themselves back sometimes — but putting yourself forward can be especially tough when you already feel like the odd one out. This session is for every person who's ever thought "I'm not ready yet" even when they were bursting with ideas. Whether you're navigating neurodivergence, underrepresentation, or just the very human experience of not seeing yourself reflected in the rooms you're trying to enter — this one's for you. Join me for an honest, inclusive discussion where I'll share strategies for what it actually takes to find your voice and give you a nudge to use it. We'll talk about the unique challenges that come with being "different" in tech spaces, the strategies that flopped, the ones that actually worked, and the practical first steps to showing up as a speaker, blogger, podcaster, content creator, or internal advocate — without waiting until you feel "ready." Spoiler: you're already ready. You just need the map — and maybe someone in your corner who believes it before you do.
General Session
The Recovering Shy Person's Guide to Putting Yourself Forward
Lenore Flower
Data Plumber, LLC
Session Goals: • Identify barriers "different" professionals face — neurodivergence, underrepresentation, and patterns that hold you back. • Apply strategies for finding your voice as a speaker, blogger, or advocate. • Define one concrete next step toward visibility you can act on today.
Session Prerequisites: • No prior knowledge required.
Track: Professional Development
Level: Level 100
-
-
-
Room: TBD
Have you ever found yourself holding back from tackling a new project, submitting that conference proposal, or publishing that blog post because the "what ifs" just became too much? Everyone holds themselves back sometimes — but putting yourself forward can be especially tough when you already feel like the odd one out. This session is for every person who's ever thought "I'm not ready yet" even when they were bursting with ideas. Whether you're navigating neurodivergence, underrepresentation, or just the very human experience of not seeing yourself reflected in the rooms you're trying to enter — this one's for you. Join me for an honest, inclusive discussion where I'll share strategies for what it actually takes to find your voice and give you a nudge to use it. We'll talk about the unique challenges that come with being "different" in tech spaces, the strategies that flopped, the ones that actually worked, and the practical first steps to showing up as a speaker, blogger, podcaster, content creator, or internal advocate — without waiting until you feel "ready." Spoiler: you're already ready. You just need the map — and maybe someone in your corner who believes it before you do. -
-
General Session
FinOps for Data: Reclaiming Control Through Cost Transparency
Ben DeBow
In today's technology landscape, every organization is developing a FinOps practice as all technology decisions carry financial implications. As data professionals, we face a striking reality: enterprise data footprints consume 20-40% of technology budgets, yet DBAs and data engineers often struggle to maintain ownership over code, data models, and architectures. The time has come to leverage our unique data expertise to change this dynamic. FinOps for data provides a framework that transforms technical objections into financial conversations everyone understands. Rather than simply stating "this is bad code," imagine the impact of declaring "this code will cost $48,999 over the next three years." Suddenly, stakeholders are listening. This session explores how data professionals can align with FinOps teams to drive better decisions around data lifecycle management and architectural choices. You'll learn strategies to quantify the financial impact of data design decisions and build a data-centric FinOps practice. We'll discuss developing metrics that communicate technical debt effectively and creating governance structures that balance innovation with cost. Join us to discover how FinOps for data can help you regain influence, drive better conversations about data architecture, and establish yourself as a strategic partner in technology investment decisions.
General Session
FinOps for Data: Reclaiming Control Through Cost Transparency
Ben DeBow
Fortified
Session Goals: • Translate technical data issues into clear financial impact with FinOps for Data. • Quantify costs of data designs, models, and code to influence stakeholders. • Build metrics and governance for cost transparency in data architecture.
Session Prerequisites: • Basic understanding of database administration or data engineering concepts. • Familiarity with SQL Server environments and common performance challenges. • Some exposure to cloud computing cost models (helpful but not required).
Track: Development
Level: Level 200
Theme: AI + Data
-
-
-
Room: TBD
In today's technology landscape, every organization is developing a FinOps practice as all technology decisions carry financial implications. As data professionals, we face a striking reality: enterprise data footprints consume 20-40% of technology budgets, yet DBAs and data engineers often struggle to maintain ownership over code, data models, and architectures. The time has come to leverage our unique data expertise to change this dynamic. FinOps for data provides a framework that transforms technical objections into financial conversations everyone understands. Rather than simply stating "this is bad code," imagine the impact of declaring "this code will cost $48,999 over the next three years." Suddenly, stakeholders are listening. This session explores how data professionals can align with FinOps teams to drive better decisions around data lifecycle management and architectural choices. You'll learn strategies to quantify the financial impact of data design decisions and build a data-centric FinOps practice. We'll discuss developing metrics that communicate technical debt effectively and creating governance structures that balance innovation with cost. Join us to discover how FinOps for data can help you regain influence, drive better conversations about data architecture, and establish yourself as a strategic partner in technology investment decisions. -
-
General Session
From Disarray to Hooray: SQL Server Functions That Simplify T-SQL
Peter Kruis
Stephanie Reis
Ever run into one of those SQL Server moments where you think: "There has to be an easier way to do this"? The kind where you've got a chain of string functions, a CASE expression that reads like a crime scene, a comment that basically says "don't touch this", and code that has been copied since 2005. In a lot of cases, we don't need those messy custom-made functions anymore, as SQL Server already has something built in that would make our life easier. In this session, we will take you through the lesser-known dungeons of SQL Server functions. Not the ones everyone memorizes but the ones you only discover after you've written the hard way and then later realize there was a cleaner option all along. We'll show how these functions can simplify data cleanup, make reporting logic easier to read, and replace bits of customer code you've been dragging around for years. To keep it fun, we'll do this in acts. Each act is a short story; a problem we keep seeing, the typical over engineered solution, and then the "wait, that exists?" moment. Expect quick demos, a few SQL memes that will feel painfully familiar, and practical takeaways you can use the next time you've stuck in a query spaghetti. If you write T-SQL and you like solutions that are simple, predictable, and easy to maintain, you'll leave with a list of function and patterns you'll actually use.
General Session
From Disarray to Hooray: SQL Server Functions That Simplify T-SQL
Peter Kruis
Kruis Database Consultancy
Stephanie Reis
AAA Washington
Session Goals: • Discover lesser-known SQL Server functions you may not have encountered before. • Identify built-in SQL Server functions that replace common overengineered T-SQL patterns. • Apply practical function patterns to clean up data and reporting logic immediately.
Session Prerequisites: • A working knowledge of basic SQL (SELECT, JOINs, WHERE, GROUP BY). • Familiary with T-SQL syntax in SQL Server.
Track: Development
Level: Level 200
Theme: Cloud + Multi-DB
-
-
-
Room: TBD
Ever run into one of those SQL Server moments where you think: "There has to be an easier way to do this"? The kind where you've got a chain of string functions, a CASE expression that reads like a crime scene, a comment that basically says "don't touch this", and code that has been copied since 2005. In a lot of cases, we don't need those messy custom-made functions anymore, as SQL Server already has something built in that would make our life easier. In this session, we will take you through the lesser-known dungeons of SQL Server functions. Not the ones everyone memorizes but the ones you only discover after you've written the hard way and then later realize there was a cleaner option all along. We'll show how these functions can simplify data cleanup, make reporting logic easier to read, and replace bits of customer code you've been dragging around for years. To keep it fun, we'll do this in acts. Each act is a short story; a problem we keep seeing, the typical over engineered solution, and then the "wait, that exists?" moment. Expect quick demos, a few SQL memes that will feel painfully familiar, and practical takeaways you can use the next time you've stuck in a query spaghetti. If you write T-SQL and you like solutions that are simple, predictable, and easy to maintain, you'll leave with a list of function and patterns you'll actually use. -
-
Lightning Talk
I Let an LLM Build My Semantic Model (Here's What Happened)
Reid Havens
I handed Claude Code the keys to a semantic model through Tabular Editor's CLI and told it to build, document, and refactor it. This is what actually happened. In this lightning talk, you'll see a real workflow where an LLM drives Tabular Editor CLI to do real semantic model development work. Writing DAX measures, organizing display folders, renaming objects, creating calculation groups, and generating documentation. You'll see how Semantic Link validates the output and catches what the LLM gets wrong. Some of it was impressive. Some of it needed guardrails. All of it was faster than doing it by hand. You'll walk away knowing what LLMs can reliably do today in semantic model development, where they confidently give you wrong answers, and how to set up a workflow that gets the most out of both.
Lightning Talk
I Let an LLM Build My Semantic Model (Here's What Happened)
Reid Havens
Analytic Endeavors
Session Goals: • See a real LLM-driven semantic model development workflow using Tabular Editor CLI. • Know what LLMs handle well in model development and where they need human oversight. • A practical setup for integrating LLMs into your own development process.
Session Prerequisites: • Experience building semantic models and writing DAX in Power BI. • No experience with LLMs, Claude Code, or Tabular Editor CLI required.
Track: Analytics
Level: Level 200
Theme: AI + Data
-
-
-
Room: TBD
I handed Claude Code the keys to a semantic model through Tabular Editor's CLI and told it to build, document, and refactor it. This is what actually happened. In this lightning talk, you'll see a real workflow where an LLM drives Tabular Editor CLI to do real semantic model development work. Writing DAX measures, organizing display folders, renaming objects, creating calculation groups, and generating documentation. You'll see how Semantic Link validates the output and catches what the LLM gets wrong. Some of it was impressive. Some of it needed guardrails. All of it was faster than doing it by hand. You'll walk away knowing what LLMs can reliably do today in semantic model development, where they confidently give you wrong answers, and how to set up a workflow that gets the most out of both. -
-
General Session
Migrating Your SQL Server DBA Skills to PostgreSQL
Dave Bland
As the industry changes, the need for Database Administrators that can support multiple database systems has increased. Of the many database systems that are available, PostgreSQL is becoming more and more popular. In this session we will cover a number of the most common DBA tasks that a SQL Server DBA completes on a regular basis and review how those same tasks can be completed in PostgreSQL. We will go over a few of the common tools using to administer a PostgreSQL database. The terminology can at times be a bit confusing, this session we will cover some the common terms in PostgreSQL and how they relate to the familiar terms in SQL Server. The concept of vacuuming is something that SQL Server DBA's are not used to working with, we will cover what it is and why it is important to manage. Permissions and security will be covered as well. If you are looking to move into the PostgreSQL world and have some experience as a SQL Server DBA, this might be a good session to give you a starting point to launch that journey.
General Session
Migrating Your SQL Server DBA Skills to PostgreSQL
Dave Bland
Lexia Learning
Session Goals: • The attendee will be able to state some common PostgreSQL terms and how they related to tasks in SQL Server. • The attendee will be able to list common tools used to manage a PostgreSQL database. • The attendee will be able to state the role of the vacuuming process in a PostgreSQL system.
Session Prerequisites: • Some basic SQL Server DBA skills would be helpful.
Track: Database Management
Level: Level 200
Theme: Cloud + Multi-DB
-
-
-
Room: TBD
As the industry changes, the need for Database Administrators that can support multiple database systems has increased. Of the many database systems that are available, PostgreSQL is becoming more and more popular. In this session we will cover a number of the most common DBA tasks that a SQL Server DBA completes on a regular basis and review how those same tasks can be completed in PostgreSQL. We will go over a few of the common tools using to administer a PostgreSQL database. The terminology can at times be a bit confusing, this session we will cover some the common terms in PostgreSQL and how they relate to the familiar terms in SQL Server. The concept of vacuuming is something that SQL Server DBA's are not used to working with, we will cover what it is and why it is important to manage. Permissions and security will be covered as well. If you are looking to move into the PostgreSQL world and have some experience as a SQL Server DBA, this might be a good session to give you a starting point to launch that journey. -
-
General Session
No Black Box: How Fabric DW Works, Performs, and Gets Monitored Right
Mariya Ali
Most Microsoft Fabric Data Warehouse users know what to do — but not what's actually happening beneath the query they just ran. In this session, you'll get a first-hand architectural walkthrough of Fabric DW from the product team that designed and shipped it. We'll start where most sessions don't — the engine itself. How does Fabric DW separate compute from storage using OneLake? What is V-Order optimization, why does it exist, and when does it hurt you instead of help you? How does the concurrency model work across SQL pools, and what are the real-world capacity implications of the choices you make today? From there, we'll move into performance: the patterns that silently kill query throughput, how to read query execution plans in a distributed columnar engine — not in theory, but in a live demo using real warehouse telemetry. Finally, we'll deep-dive into the monitoring and observability stack: Query Insights, Warehouse Insights, SQL Pool Insights, and Query Plan Visualization. You'll leave knowing exactly which DMV-equivalent views surface the signals that matter, how to build proactive alerting, and how enterprise teams — including large healthcare organizations running production workloads — are using these tools to drive meaningful cost reduction and SLA reliability. Whether you're evaluating Fabric, mid-migration, or already in production and hitting walls you don't fully understand, this session gives you the architecture mental model you've been missing.
General Session
No Black Box: How Fabric DW Works, Performs, and Gets Monitored Right
Mariya Ali
Microsoft
Session Goals: • Understand how Fabric DW's architecture drives query performance decisions. • Navigate the full monitoring stack to diagnose production issues fast. • Leave with an observability playbook you can apply on Monday.
Session Prerequisites: • Attendees should have working familiarity with SQL and basic data warehousing concepts (tables, schemas, query execution). • Prior exposure to Microsoft Fabric is helpful but not required.
Track: Architecture
Level: Level 300
Theme: AI + Data
-
-
-
Room: TBD
Most Microsoft Fabric Data Warehouse users know what to do — but not what's actually happening beneath the query they just ran. In this session, you'll get a first-hand architectural walkthrough of Fabric DW from the product team that designed and shipped it. We'll start where most sessions don't — the engine itself. How does Fabric DW separate compute from storage using OneLake? What is V-Order optimization, why does it exist, and when does it hurt you instead of help you? How does the concurrency model work across SQL pools, and what are the real-world capacity implications of the choices you make today? From there, we'll move into performance: the patterns that silently kill query throughput, how to read query execution plans in a distributed columnar engine — not in theory, but in a live demo using real warehouse telemetry. Finally, we'll deep-dive into the monitoring and observability stack: Query Insights, Warehouse Insights, SQL Pool Insights, and Query Plan Visualization. You'll leave knowing exactly which DMV-equivalent views surface the signals that matter, how to build proactive alerting, and how enterprise teams — including large healthcare organizations running production workloads — are using these tools to drive meaningful cost reduction and SLA reliability. Whether you're evaluating Fabric, mid-migration, or already in production and hitting walls you don't fully understand, this session gives you the architecture mental model you've been missing. -
-
General Session
Choosing the Right Data Store: An Overview of Azure Data Platform Choices
Denny Cherry
Choosing the right data store can make or break the performance and costs of your application. In this session, you will learn about all the options available to you in the Azure Data Platform. You'll learn about CosmosDB, Azure SQL Database, Azure Postgres, and MySQL services, as well as Databricks and HDInsight. You will also learn how to leverage caching to reduce the overall workload on your data tier. We will discuss the notion of polyglot persistence, storing different data in different data stores with the goal of optimizing your overall data architecture, for flexibility and performance.
General Session
Choosing the Right Data Store: An Overview of Azure Data Platform Choices
Denny Cherry
Denny Cherry & Associates Consulting
Session Goals: • How to select the correct database platform for your database. • Differences between the different Azure offerings.
Session Prerequisites: • None.
Track: Architecture
Level: Level 200
Theme: AI + Data
-
-
-
Room: TBD
Choosing the right data store can make or break the performance and costs of your application. In this session, you will learn about all the options available to you in the Azure Data Platform. You'll learn about CosmosDB, Azure SQL Database, Azure Postgres, and MySQL services, as well as Databricks and HDInsight. You will also learn how to leverage caching to reduce the overall workload on your data tier. We will discuss the notion of polyglot persistence, storing different data in different data stores with the goal of optimizing your overall data architecture, for flexibility and performance. -
-
General Session
AI Coding Agents for Power BI: From Hype to Daily Workflow
Garett Medlin
You've heard that AI can write DAX. Can it also review a measure for performance anti-patterns, assess an entire semantic model's health, identify unused fields across report pages, and execute an optimization workflow? This session is a live, demo-driven walkthrough of what an AI-augmented Power BI development workflow actually looks like in 2026. You'll see an AI coding agent connected to a live semantic model, authoring and reviewing measures in TMDL format, running model health assessments, and catching issues that take hours to find manually. Every concept is demonstrated against real models, and every pattern shown is something you can build into your own workflow. You'll leave knowing what AI agents can actually do with your Power BI models today, how they connect to your data platform through tools like MCP and the Fabric CLI, and where they still need a human in the loop.
General Session
AI Coding Agents for Power BI: From Hype to Daily Workflow
Garett Medlin
P3 Adaptive
Session Goals: • What AI agents can do with Power BI models today: measure authoring, health assessment, DAX review, field analysis, demonstrated live. • How agents connect via MCP, Fabric CLI, and Tabular Editor. • Where agents excel and where they fail, so you set realistic expectations.
Session Prerequisites: • Working knowledge of Power BI semantic models (measures, relationships, DAX basics). • Familiarity with any code editor. • No prior experience with AI coding agents, MCP, or TMDL required.
Track: Analytics
Level: Level 200
Theme: AI + Data
-
-
-
Room: TBD
You've heard that AI can write DAX. Can it also review a measure for performance anti-patterns, assess an entire semantic model's health, identify unused fields across report pages, and execute an optimization workflow? This session is a live, demo-driven walkthrough of what an AI-augmented Power BI development workflow actually looks like in 2026. You'll see an AI coding agent connected to a live semantic model, authoring and reviewing measures in TMDL format, running model health assessments, and catching issues that take hours to find manually. Every concept is demonstrated against real models, and every pattern shown is something you can build into your own workflow. You'll leave knowing what AI agents can actually do with your Power BI models today, how they connect to your data platform through tools like MCP and the Fabric CLI, and where they still need a human in the loop. -
-
General Session
Advanced Data Recovery Techniques
Paul Randal
Disasters happen – plain and simple. When disaster strikes a database you're responsible for, and backups and repair fail, how can you salvage data, and possibly your company and your job? This is where advanced data recovery techniques come in. Using undocumented tools and deep knowledge of database structures, you can manually patch up the database enough to extract critical data. This demo-heavy session will show you methods I've used extensively to salvage data for real-life clients after catastrophic corruption. You won't believe what it's possible to do!
General Session
Advanced Data Recovery Techniques
Paul Randal
SQLskills
Session Goals: • Show how databases can be easily corrupted to allow DR testing. • Show how lack of backups doesn't always mean total data loss. • Show how deep the rabbit hole goes…
Session Prerequisites: • Familiarity with terms/concepts like DBCC CHECKDB, pages, records, locks, indexes. Although a 400-level session, it doesn't require that level of knowledge.
Track: Database Management
Level: Level 400
Theme: AI + Data
-
-
-
Room: TBD
Disasters happen – plain and simple. When disaster strikes a database you're responsible for, and backups and repair fail, how can you salvage data, and possibly your company and your job? This is where advanced data recovery techniques come in. Using undocumented tools and deep knowledge of database structures, you can manually patch up the database enough to extract critical data. This demo-heavy session will show you methods I've used extensively to salvage data for real-life clients after catastrophic corruption. You won't believe what it's possible to do! -
-
Learning Pathway
Adapting for AI Readiness as a Data Professional | AI in the SQL World: Tools We Have & Tips for Using Them
Taiob Ali
In the first session of this learning pathway, we walked through the tools, technologies, and skills needed for a data professional to up their game in our AI-impacted world. In this session, we'll focus on what to do as Microsoft, and our own managers, become increasingly focused on having us integrate AI into our daily workflows, what does that actually mean for us as data professionals that work on Microsoft data platforms? If our boss instructs us to do X% of work "with AI", what does that mean and how can we do that without negatively impacting our work? Maybe we've heard our developer friends going on and on about how much coding assistants have helped them and we want to explore what GitHub Copilot in SSMS can do for us. It could be that we're being asked to support "AI search" in our database – but we're not quite sure what that means and how to do that. If any of those scenarios (or similar) resonate with you, join us for this session as we explore the AI-related tools and capabilities in the Microsoft SQL ecosystem and how you can use them to boost productivity, support new workloads, and maybe even have a little more fun at work as well.
Learning Pathway
Adapting for AI Readiness as a Data Professional | AI in the SQL World: Tools We Have & Tips for Using Them
Taiob Ali
GMO LLC
Session Goals: • Learn about helpful ways to integrate AI into our daily workflows. • Learn about what coding assistants are and how they can helpful in data-related development. • Learn about implementing "AI search" in Azure SQL and SQL Server and what developers mean when they ask us to do that.
Session Prerequisites: • A basic knowledge of SQL is preferred.
Track: Analytics
Level: Level 200
Theme: AI + Data
-
-
-
Room: TBD
In the first session of this learning pathway, we walked through the tools, technologies, and skills needed for a data professional to up their game in our AI-impacted world. In this session, we'll focus on what to do as Microsoft, and our own managers, become increasingly focused on having us integrate AI into our daily workflows, what does that actually mean for us as data professionals that work on Microsoft data platforms? If our boss instructs us to do X% of work "with AI", what does that mean and how can we do that without negatively impacting our work? Maybe we've heard our developer friends going on and on about how much coding assistants have helped them and we want to explore what GitHub Copilot in SSMS can do for us. It could be that we're being asked to support "AI search" in our database – but we're not quite sure what that means and how to do that. If any of those scenarios (or similar) resonate with you, join us for this session as we explore the AI-related tools and capabilities in the Microsoft SQL ecosystem and how you can use them to boost productivity, support new workloads, and maybe even have a little more fun at work as well. -
-
General Session
Forget Visual Query Plans, Parse the XML
Richard Douglas
The SSMS graphical execution plan is a simplification. It hides thread skew in parallel operators, buries memory grant wait times, shows only the first missing index when there are several, silently misorders index column recommendations, and renders scalar UDFs as zero-cost operations. The XML underneath shows all — and most people never read it. In this session, Richard demonstrates what the graphical plan isn't showing you, using a live demo of Plan Investigator — his free, public beta built specifically to surface these hidden details. We'll parse real execution plans and expose what SSMS misses: per-thread row distribution in parallel plans, memory queuing time before a query could even start executing, composite index column gaps and much much more! Every finding is traced back to the specific XML attribute that contains it, attendees will leave understanding not just what the tool found, but how to find it themselves using nothing but native SQL Server capabilities.
General Session
Forget Visual Query Plans, Parse the XML
Richard Douglas
Redgate
Session Goals: • Identify what SSMS hides in the graphical plan and where to find it in the XML. • Diagnose parallel skew, memory grant issues, index column gaps, and more from the XML. • Leave with Plan Investigator — a free public beta that highlights these issues instantly.
Session Prerequisites: • Working knowledge of SQL Server execution plans. • You should be able to read a basic graphical plan in SSMS. • Familiarity with index seeks, sorts, & parallelism is helpful but not required. • Suitable for DBAs & developers who tune queries regularly.
Track: Development
Level: Level 300
Theme: AI + Data
-
-
-
Room: TBD
The SSMS graphical execution plan is a simplification. It hides thread skew in parallel operators, buries memory grant wait times, shows only the first missing index when there are several, silently misorders index column recommendations, and renders scalar UDFs as zero-cost operations. The XML underneath shows all — and most people never read it. In this session, Richard demonstrates what the graphical plan isn't showing you, using a live demo of Plan Investigator — his free, public beta built specifically to surface these hidden details. We'll parse real execution plans and expose what SSMS misses: per-thread row distribution in parallel plans, memory queuing time before a query could even start executing, composite index column gaps and much much more! Every finding is traced back to the specific XML attribute that contains it, attendees will leave understanding not just what the tool found, but how to find it themselves using nothing but native SQL Server capabilities. -
-
Deep Dive
How to Properly Handle LOB Data in SQL Server
Uwe Ricken
Larger-than-life data, often referred to as Large Object (LOB) data, such as JSON, XML, documents, images has become an integral part of modern database systems. Microsoft SQL Server offers powerful features and strategies for effectively managing LOB data. This session will take you on a journey through the intricate world of LOB data in SQL Server, offering insights, best practices, and practical techniques to streamline your data management strategies.
Deep Dive
How to Properly Handle LOB Data in SQL Server
Uwe Ricken
db Berater GmbH
Session Goals: • Understand SQL Server’s internal handling of LOB data and how it affects performance. • Learn advanced techniques for designing, storing, and optimizing LOB‑heavy workloads. • Apply best practices and real‑world strategies to manage, streamline, and troubleshoot large‑scale LOB data environments.
Session Prerequisites: • Strong SQL Server administration skills, including storage architecture, and performance tuning. • Familiarity with data structures and data types is expected. • Experience with complex data‑management workloads will help for this session.
Track: Database Management
Level: Level 300
Theme: AI + Data
-
-
-
Room: TBD
Larger-than-life data, often referred to as Large Object (LOB) data, such as JSON, XML, documents, images has become an integral part of modern database systems. Microsoft SQL Server offers powerful features and strategies for effectively managing LOB data. This session will take you on a journey through the intricate world of LOB data in SQL Server, offering insights, best practices, and practical techniques to streamline your data management strategies. -
-
Lightning Talk
DIY Log Shipping for RDS
Matthew Gantz
Log shipping is a powerful and reliable tool for migrating databases and disaster recovery. Even with the rise of Availability Groups and other modern features, its simplicity and predictability make it an essential part of the DBA toolkit. Unfortunately, AWS RDS does not support log shipping out of the box. In this session, I will show you how to build your own lightweight solution using a small set of scripts. We will walk through a practical approach for migrating a database from an on-premises SQL Server instance to AWS RDS with minimal downtime and a simple, low-risk rollback strategy. If you are planning a cloud migration or want to use RDS as a DR solution, this this talk will give you a proven approach you can implement right away.
Lightning Talk
DIY Log Shipping for RDS
Matthew Gantz
Versant
Session Goals: • Be able to setup a log shipping solution from on-prem to AWS. • Be able to reverse the log shipping to rollback to on-prem.
Session Prerequisites: • Some background knowledge of log shipping is helpful but not strictly required.
Track: Database Management
Level: Level 200
Theme: Cloud + Multi-DB
-
-
-
Room: TBD
Log shipping is a powerful and reliable tool for migrating databases and disaster recovery. Even with the rise of Availability Groups and other modern features, its simplicity and predictability make it an essential part of the DBA toolkit. Unfortunately, AWS RDS does not support log shipping out of the box. In this session, I will show you how to build your own lightweight solution using a small set of scripts. We will walk through a practical approach for migrating a database from an on-premises SQL Server instance to AWS RDS with minimal downtime and a simple, low-risk rollback strategy. If you are planning a cloud migration or want to use RDS as a DR solution, this this talk will give you a proven approach you can implement right away. -
-
General Session
SQL Server Always On Deployments: What Works, What Breaks, and What to Do
Marcelo Adade
Always On Availability Groups and Failover Cluster Instances are powerful features, but real-world deployments often come with challenges that go far beyond the basics. In this demo-packed session, we’ll share practical lessons learned from administering SQL Server Always On environments, focusing on what actually works in production, what commonly breaks, and how to respond effectively. We’ll explore configuration decisions, recurring pitfalls, troubleshooting patterns, and operational realities that are not always obvious from the documentation. Through real examples and practical demonstrations, you’ll gain insight into how to better plan, validate, and support Always On deployments across different scenarios. By the end of the session, attendees will be better prepared to design, operate, and troubleshoot the majority of Always On implementations they encounter.
General Session
SQL Server Always On Deployments: What Works, What Breaks, and What to Do
Marcelo Adade
DBBITS
Session Goals: • Plan reliable Always On AG and FCI deployments. • Identify and prevent common configuration and operational failures. • Troubleshoot real-world Always On issues more effectively.
Session Prerequisites: • Some knowledge and experience related to HADR and Always On AGs and FCIs in SQL Server.
Track: Database Management
Level: Level 300
Theme: Cloud + Multi-DB
-
-
-
Room: TBD
Always On Availability Groups and Failover Cluster Instances are powerful features, but real-world deployments often come with challenges that go far beyond the basics. In this demo-packed session, we’ll share practical lessons learned from administering SQL Server Always On environments, focusing on what actually works in production, what commonly breaks, and how to respond effectively. We’ll explore configuration decisions, recurring pitfalls, troubleshooting patterns, and operational realities that are not always obvious from the documentation. Through real examples and practical demonstrations, you’ll gain insight into how to better plan, validate, and support Always On deployments across different scenarios. By the end of the session, attendees will be better prepared to design, operate, and troubleshoot the majority of Always On implementations they encounter. -
-
General Session
Secure CI/CD for Power BI with Github
Akshata Revankar
As Power BI becomes mission-critical, manual deployments create risk: inconsistent environments, accidental overwrites, and limited traceability. This session demonstrates how to implement secure, repeatable CI/CD for Power BI using Github repos for CI and GitHub Actions for CD. Attendees will learn how to structure workspaces, version Power BI assets, automate deployments, and enforce governance. Real-world examples show how organizations reduce errors, improve reliability, and accelerate delivery by aligning Power BI with modern DevOps practices. Key Topics Covered Challenges in traditional Power BI lifecycle management Reference architecture for CI/CD across dev/test/prod Using deployment pipelines for controlled environment promotion Versioning PBIP files and integrating with GitHub Automating deployments with GitHub Actions and service principals Governance, auditability, and policy-as-code considerations
General Session
Secure CI/CD for Power BI with Github
Akshata Revankar
Microsoft
Session Goals: • Learn how to implement secure, automated CI/CD pipelines for Power BI using GitHub. • Understand versioning and deployment best practices that improve reliability and governance.
Session Prerequisites: • Familiarity with Power BI and basic DevOps concepts recommended. • Some exposure to GitHub helpful.
Track: Development
Level: Level 300
Theme: AI + Data
-
-
-
Room: TBD
As Power BI becomes mission-critical, manual deployments create risk: inconsistent environments, accidental overwrites, and limited traceability. This session demonstrates how to implement secure, repeatable CI/CD for Power BI using Github repos for CI and GitHub Actions for CD. Attendees will learn how to structure workspaces, version Power BI assets, automate deployments, and enforce governance. Real-world examples show how organizations reduce errors, improve reliability, and accelerate delivery by aligning Power BI with modern DevOps practices. Key Topics Covered Challenges in traditional Power BI lifecycle management Reference architecture for CI/CD across dev/test/prod Using deployment pipelines for controlled environment promotion Versioning PBIP files and integrating with GitHub Automating deployments with GitHub Actions and service principals Governance, auditability, and policy-as-code considerations -
-
General Session
AI-Assisted BI (AIBI): Using LLMs to Build and Maintain Semantic Models
Reid Havens
LLMs are changing how BI developers work, and most teams are barely scratching the surface. This session goes beyond asking Copilot to write a measure. We'll look at how tools like Claude Code and the Power BI MCP server can be used to author DAX, document models, refactor naming conventions, generate testing notebooks, and enforce standards across workspaces. You'll see real workflows for using LLMs to build and maintain semantic models in Fabric, not hypothetical demos, but the way this actually works in day-to-day development. We'll cover prompting strategies that produce reliable DAX, how to validate LLM output against your model metadata with Semantic Link, and where LLMs still fall short so you know when to step in. Whether you're a solo developer looking to move faster or part of a team trying to standardize model quality, this session gives you a practical starting point for integrating LLMs into your BI development workflow.
General Session
AI-Assisted BI (AIBI): Using LLMs to Build and Maintain Semantic Models
Reid Havens
Analytic Endeavors
Session Goals: • Know where LLMs add real value in semantic model development and where they don't. • Practical prompting and validation patterns for DAX authoring and model refactoring. • A workflow for integrating LLMs into your existing BI development process.
Session Prerequisites: • Comfortable building semantic models in Power BI and writing DAX measures. • No prior experience with LLMs, Claude Code, or the MCP server is needed. • Familiarity with Fabric workspaces is helpful.
Track: Analytics
Level: Level 300
Theme: AI + Data
-
-
-
Room: TBD
LLMs are changing how BI developers work, and most teams are barely scratching the surface. This session goes beyond asking Copilot to write a measure. We'll look at how tools like Claude Code and the Power BI MCP server can be used to author DAX, document models, refactor naming conventions, generate testing notebooks, and enforce standards across workspaces. You'll see real workflows for using LLMs to build and maintain semantic models in Fabric, not hypothetical demos, but the way this actually works in day-to-day development. We'll cover prompting strategies that produce reliable DAX, how to validate LLM output against your model metadata with Semantic Link, and where LLMs still fall short so you know when to step in. Whether you're a solo developer looking to move faster or part of a team trying to standardize model quality, this session gives you a practical starting point for integrating LLMs into your BI development workflow. -
-
General Session
SQL Server Versioning Explained
Bob Ward
Version store, Accelerated Database Recovery (ADR), Optimized Locking (OL), and RSCI. What do these all have in common? Versioning. In this session the mysteries of SQL Server versioning will be unlocked. You will learn the internals of how it all works, what to use and when, and examples including how to debug it. And we will bring the numbers so you will see exactly how your SQL workload will be affected by using versioning. You will walk aways not just learning how things work but details to help you decide whether you should use options like ADR or OL.
General Session
SQL Server Versioning Explained
Bob Ward
Microsoft
Session Goals: • Understand how versioning works in SQL Server. • Learn how using various database options control versioning and affect performance. • Bust myths about how versioning can negatively affect your application.
Session Prerequisites: • Fundamental understanding of SQL Server.
Track: Database Management
Level: Level 300
Theme: Cloud + Multi-DB
-
-
-
Room: TBD
Version store, Accelerated Database Recovery (ADR), Optimized Locking (OL), and RSCI. What do these all have in common? Versioning. In this session the mysteries of SQL Server versioning will be unlocked. You will learn the internals of how it all works, what to use and when, and examples including how to debug it. And we will bring the numbers so you will see exactly how your SQL workload will be affected by using versioning. You will walk aways not just learning how things work but details to help you decide whether you should use options like ADR or OL. -
-
General Session
Shattering Records: Architecting a 1,000,000 IOPS SQL Server
Christoph Petersen
Chris Madden
Modern SQL Server workloads, from massive parallel processing to high-concurrency OLTP, now demand I/O performance once limited to specialized on-premises hardware. Achieving 1,000,000 IOPS in cloud or virtual environments is no longer theoretical—it is a requirement for scale. This session deconstructs the infrastructure needed to sustain this level of performance. We will analyze the convergence of compute, storage, and networking, identifying where standard configurations fail under extreme load. We’ll dissect technical hurdles including CPU selection, NUMA-aware memory association, and networking throughput limits. Additionally, we map these infrastructure layers back to SQL Server internals to tune the database engine to leverage the performance provided by the underlying hardware. Designed for DBAs and architects pushing physical limits, this session provides a technical blueprint for the fastest modern database environments. Attendees will leave with: * Hardware Blueprint: Comparing CPU families and storage solutions for sub-millisecond latency. * Configuration Checklists: Methodologies for host-level storage offloading and networking. * Tuning Guide: SQL Server changes to prevent engine bottlenecks. Cost-Performance Analysis: Frameworks for right-sizing resources without over-provisioning.
General Session
Shattering Records: Architecting a 1,000,000 IOPS SQL Server
Christoph Petersen
Google
Chris Madden
Google Cloud
Session Goals: • Performance: Aligning VM specifications with next-gen block storage to eliminate I/O bottlenecks. • Continuity: Evaluate HA/DR patterns that maintain performance parity during failover. • Cost: Balance resource elasticity and SQL Server core licensing to maximize price-performance.
Session Prerequisites: • Designed for DBAs and architects who need to push SQL Server to its physical limits. • This session provides the technical blueprint for building one of the fastest database environments possible today.
Track: Architecture
Level: Level 300
Theme: Cloud + Multi-DB
-
-
-
Room: TBD
Modern SQL Server workloads, from massive parallel processing to high-concurrency OLTP, now demand I/O performance once limited to specialized on-premises hardware. Achieving 1,000,000 IOPS in cloud or virtual environments is no longer theoretical—it is a requirement for scale. This session deconstructs the infrastructure needed to sustain this level of performance. We will analyze the convergence of compute, storage, and networking, identifying where standard configurations fail under extreme load. We’ll dissect technical hurdles including CPU selection, NUMA-aware memory association, and networking throughput limits. Additionally, we map these infrastructure layers back to SQL Server internals to tune the database engine to leverage the performance provided by the underlying hardware. Designed for DBAs and architects pushing physical limits, this session provides a technical blueprint for the fastest modern database environments. Attendees will leave with: * Hardware Blueprint: Comparing CPU families and storage solutions for sub-millisecond latency. * Configuration Checklists: Methodologies for host-level storage offloading and networking. * Tuning Guide: SQL Server changes to prevent engine bottlenecks. Cost-Performance Analysis: Frameworks for right-sizing resources without over-provisioning. -
-
General Session
Mastering Fabric Security: Safeguarding Your Analytics
Bradley Schacht
Security is at the heart of Microsoft Fabric and there is no shortage of options. Virtual networks, private links, outbound access protection, conditional access and more. Let's not forget the new kid on the block, OneLake security. We will walk through a live demo on how to use VNETs, private endpoints, and workspace private link to reduce exposure; apply workspace permissions and item-level sharing for least privileged data access; and implement RLS/CLS in lakehouse and warehouse for segregation of duties. We’ll also cover OneLake security to ensure your data remains protected across the entire Fabric ecosystem. By the end of this session, you will be well versed in the options, when to use each, and how to layer the options for maximum protection.
General Session
Mastering Fabric Security: Safeguarding Your Analytics
Bradley Schacht
Microsoft
Session Goals: • Know all the options for securing a Fabric workspace and the items within it. • Know how each option works in isolation. • How they can work together to a multi-layered security approach.
Session Prerequisites: • A general understanding of Microsoft Fabric. • Know the basics of how data is stored in OneLake. • Basic knowledge of Azure networking is advised but not required.
Track: Analytics
Level: Level 200
Theme: Security
-
-
-
Room: TBD
Security is at the heart of Microsoft Fabric and there is no shortage of options. Virtual networks, private links, outbound access protection, conditional access and more. Let's not forget the new kid on the block, OneLake security. We will walk through a live demo on how to use VNETs, private endpoints, and workspace private link to reduce exposure; apply workspace permissions and item-level sharing for least privileged data access; and implement RLS/CLS in lakehouse and warehouse for segregation of duties. We’ll also cover OneLake security to ensure your data remains protected across the entire Fabric ecosystem. By the end of this session, you will be well versed in the options, when to use each, and how to layer the options for maximum protection. -
-
General Session
The Fellowship of the Star Schema: Transforming OLTP Data for Power BI
Kristyna Ferris
One does not simply build reports on OLTP data. Join us on an epic journey from the depths of raw, normalized tables to the shining halls of a well-modeled star schema fit for Power BI greatness. We will demystify the differences between OLTP and OLAP, walk through star schema vs snowflake (and why Power BI has a clear favorite), and show you the compression differences between normalized and columnar storage that will make you never look at an OLTP model the same way again. We will cover the different types of dimensions, when to use them, and how to handle Slowly Changing Dimensions without losing your mind. You will leave knowing exactly how to identify a bad model, how to convert OLTP data into a clean star schema, and how to forge relationships that rule them all. Whether you are a data wizard, a curious hobbit, or somewhere in between, your models will never be the same.
General Session
The Fellowship of the Star Schema: Transforming OLTP Data for Power BI
Kristyna Ferris
P3 Adaptive
Session Goals: • Difference between OLAP and OLTP • How to create a star schema from scratch • What a slowly changing dimension is and when to use it
Session Prerequisites: • None! Come and learn!
Track: Analytics
Level: Level 100
Theme: AI + Data
-
-
-
Room: TBD
One does not simply build reports on OLTP data. Join us on an epic journey from the depths of raw, normalized tables to the shining halls of a well-modeled star schema fit for Power BI greatness. We will demystify the differences between OLTP and OLAP, walk through star schema vs snowflake (and why Power BI has a clear favorite), and show you the compression differences between normalized and columnar storage that will make you never look at an OLTP model the same way again. We will cover the different types of dimensions, when to use them, and how to handle Slowly Changing Dimensions without losing your mind. You will leave knowing exactly how to identify a bad model, how to convert OLTP data into a clean star schema, and how to forge relationships that rule them all. Whether you are a data wizard, a curious hobbit, or somewhere in between, your models will never be the same. -
-
Deep Dive
Stop AI Hallucinations: Grounding Agents with Fabric Semantic Models
Sajitha Sritharan
Your AI agent confidently reports Q3 revenue as $47 million—but the real number is $74 million. The math checks out. The meaning doesn’t. This is the grounding problem, and it’s the fastest way to lose executive trust in enterprise AI. Most teams try to fix this with better prompts or more powerful models. That approach fails because the root issue isn’t intelligence—it’s context. AI systems lack a shared, governed understanding of business data. In this session, you’ll learn how to solve grounding using Microsoft Fabric’s layered approach. We’ll walk through three critical components: semantic models that define trusted business metrics, ontologies that align business language with data, and agent-level guardrails that constrain AI behavior. You’ll see a live, end-to-end demo of a Data Agent that initially hallucinates and produces misleading insights. Then, step by step, we’ll ground it—transforming it into a reliable, auditable system that business stakeholders can trust. This session is designed for data engineers, analytics leaders, and architects responsible for deploying AI in production environments. If you’re building—or being asked to build—AI agents on top of enterprise data, this session will give you the practical patterns to do it safely and correctly.
Deep Dive
Stop AI Hallucinations: Grounding Agents with Fabric Semantic Models
Sajitha Sritharan
Yazaki
Session Goals: • Understand why AI hallucinations occur in enterprise data scenarios. • Learn how to ground AI agents using semantic models, ontologies, and guardrails in Microsoft Fabric. • Gain practical patterns to build trustworthy, auditable AI data agents.
Session Prerequisites: • Basic understanding of data platforms, analytics concepts (e.g., metrics, models). • Familiarity with AI/LLM concepts is helpful but not required.
Track: Analytics
Level: Level 300
Theme: AI + Data
-
-
-
Room: TBD
Your AI agent confidently reports Q3 revenue as $47 million—but the real number is $74 million. The math checks out. The meaning doesn’t. This is the grounding problem, and it’s the fastest way to lose executive trust in enterprise AI. Most teams try to fix this with better prompts or more powerful models. That approach fails because the root issue isn’t intelligence—it’s context. AI systems lack a shared, governed understanding of business data. In this session, you’ll learn how to solve grounding using Microsoft Fabric’s layered approach. We’ll walk through three critical components: semantic models that define trusted business metrics, ontologies that align business language with data, and agent-level guardrails that constrain AI behavior. You’ll see a live, end-to-end demo of a Data Agent that initially hallucinates and produces misleading insights. Then, step by step, we’ll ground it—transforming it into a reliable, auditable system that business stakeholders can trust. This session is designed for data engineers, analytics leaders, and architects responsible for deploying AI in production environments. If you’re building—or being asked to build—AI agents on top of enterprise data, this session will give you the practical patterns to do it safely and correctly. -
-
Learning Pathway
A GIT Journey from Foundations to Automation for Data Professionals
Amy Abel
Michael D'Spain
Robert Volk
Most Git sessions teach generic commands but ignore the unique complexities of data environments. In this interactive session, three experts break down the Git ecosystem into a customized journey. The Journey is split into three core modules: Module 1: The Foundation (The "Why" and the "How") We start by demystifying the ecosystem. We’ll clarify the difference between Git (the engine) and GitHub (the garage). You’ll master the essential command-line basics and learn how to supercharge your productivity with Git Aliases, turning long, complex strings into short, data-friendly shortcuts. Module 2: The Branching Strategy (The "Safety Net"). We examine Branching and Merging. You’ll learn how to manage concurrent workstreams without the fear of overwriting a colleague's stored procedure or breaking the production build. Module 3: The Automation Engine (The "Future") Finally, we put your Git knowledge to work. We explore the powerful features of GitHub Actions and Webhooks, and the "hidden gems" of platform secrets that make Git feel like magic. We move beyond the basics into high-impact automation. You will learn how to: -Use GitHub Actions to handle your database deployments. -Templates to spin up perfect, pre-configured SQL projects in seconds. -use Issue Forms to require users to provide the exact info you need for database requests. Discover UI "Easter Eggs" like the instant web-editor and Webhooks that turn you into the most efficient person on your team.
Learning Pathway
A GIT Journey from Foundations to Automation for Data Professionals
Amy Abel
Straight Path
Michael D'Spain
Walmart Vizio
Robert Volk
Rob Volk
Session Goals: • Understanding the Git ecosystem, the difference between Git and GitHub. • Using branching and merging as a safety net for isolating change. • Leveraging GitHub Actions, Webhooks, Issue Templates, and automation.
Session Prerequisites: • Basic Tech Skills Familiarity with files, folders, and the command line. • DevOps Curiosity A desire to automate manual database deployments. • No Prior Git Experience Required We start with fundamentals then advance to automation.
Track: Development
Level: Level 200
Theme: Cloud + Multi-DB
-
-
-
Room: TBD
Most Git sessions teach generic commands but ignore the unique complexities of data environments. In this interactive session, three experts break down the Git ecosystem into a customized journey. The Journey is split into three core modules: Module 1: The Foundation (The "Why" and the "How") We start by demystifying the ecosystem. We’ll clarify the difference between Git (the engine) and GitHub (the garage). You’ll master the essential command-line basics and learn how to supercharge your productivity with Git Aliases, turning long, complex strings into short, data-friendly shortcuts. Module 2: The Branching Strategy (The "Safety Net"). We examine Branching and Merging. You’ll learn how to manage concurrent workstreams without the fear of overwriting a colleague's stored procedure or breaking the production build. Module 3: The Automation Engine (The "Future") Finally, we put your Git knowledge to work. We explore the powerful features of GitHub Actions and Webhooks, and the "hidden gems" of platform secrets that make Git feel like magic. We move beyond the basics into high-impact automation. You will learn how to: -Use GitHub Actions to handle your database deployments. -Templates to spin up perfect, pre-configured SQL projects in seconds. -use Issue Forms to require users to provide the exact info you need for database requests. Discover UI "Easter Eggs" like the instant web-editor and Webhooks that turn you into the most efficient person on your team. -
-
General Session
Thriving in a Tough Job Market: Résumés, Interviews, and Career Strategy
Paresh Motiwala
In a challenging economy, job hunting can feel overwhelming, but the right strategy can help you stand out even when opportunities are scarce. This highly interactive session gives IT professionals a practical, no‑nonsense roadmap for navigating the job market with confidence. We’ll break down what hiring managers look for today, how applicant‑tracking systems filter résumés, and how to position your experience so it rises to the top of the stack. You’ll learn how to craft a résumé that highlights impact instead of tasks, uses metrics to tell your story, and avoids the common mistakes that quietly get candidates screened out. We’ll also cover how to tailor your résumé for different roles without rewriting it from scratch. From there, we’ll shift to interviewing, including how to prepare for behavioral questions, how to communicate your value clearly, and how to handle tough conversations about layoffs, gaps, or career changes. We’ll also discuss strategies for staying resilient, networking effectively, and creating opportunities even when the market feels slow. Whether you’re actively searching or simply preparing for what’s next, this session will equip you with practical tools to compete and win in a difficult job market.
General Session
Thriving in a Tough Job Market: Résumés, Interviews, and Career Strategy
Paresh Motiwala
Harris County, TX
Session Goals: • Craft impact‑focused résumés that pass both human and ATS review. • Communicate value confidently and handle tough interview questions. • Use practical networking and job‑search strategies to stay competitive in a slow market.
Session Prerequisites: • No prior knowledge needed.
Track: Professional Development
Level: Level 100
-
-
-
Room: TBD
In a challenging economy, job hunting can feel overwhelming, but the right strategy can help you stand out even when opportunities are scarce. This highly interactive session gives IT professionals a practical, no‑nonsense roadmap for navigating the job market with confidence. We’ll break down what hiring managers look for today, how applicant‑tracking systems filter résumés, and how to position your experience so it rises to the top of the stack. You’ll learn how to craft a résumé that highlights impact instead of tasks, uses metrics to tell your story, and avoids the common mistakes that quietly get candidates screened out. We’ll also cover how to tailor your résumé for different roles without rewriting it from scratch. From there, we’ll shift to interviewing, including how to prepare for behavioral questions, how to communicate your value clearly, and how to handle tough conversations about layoffs, gaps, or career changes. We’ll also discuss strategies for staying resilient, networking effectively, and creating opportunities even when the market feels slow. Whether you’re actively searching or simply preparing for what’s next, this session will equip you with practical tools to compete and win in a difficult job market. -
-
General Session
AI-Powered SQL Dev with VS Code and Copilot
Anupama Natarajan
Bhaskar Natarajan
Modern database development is shifting from manual scripting to AI-assisted engineering. Using Visual Studio Code, MSSQL extensions, SQL Database Projects, and GitHub Copilot, this session demonstrates a complete AI-native workflow. You’ll learn how to generate queries, refactor stored procedures, manage schema changes, and implement CI/CD pipelines—all with Copilot assistance. We’ll also explore how AI accelerates code reviews, documentation, and debugging. This is a practical, hands-on session for developers who want to dramatically boost productivity.
General Session
AI-Powered SQL Dev with VS Code and Copilot
Anupama Natarajan
Pearl Innovations Limited
Bhaskar Natarajan
Pearl Innovations Limited
Session Goals: • Set up modern SQL development workflow in VS Code. • Use Copilot for query generation and optimization. • Implement Git-based database DevOps.
Session Prerequisites: • Basic SQL development experience. • Familiarity with Git/version control. • Optional: Exposure to DevOps practices.
Track: Database Management
Level: Level 300
Theme: AI + Data
-
-
-
Room: TBD
Modern database development is shifting from manual scripting to AI-assisted engineering. Using Visual Studio Code, MSSQL extensions, SQL Database Projects, and GitHub Copilot, this session demonstrates a complete AI-native workflow. You’ll learn how to generate queries, refactor stored procedures, manage schema changes, and implement CI/CD pipelines—all with Copilot assistance. We’ll also explore how AI accelerates code reviews, documentation, and debugging. This is a practical, hands-on session for developers who want to dramatically boost productivity. -
-
General Session
Learning to Present and Becoming Good At It
James Serra
Have you been thinking about presenting at a user group? Are you being asked to present at your work? Is learning to present one of the keys to advancing your career? Or do you just think it would be fun to present but you are too nervous to try it? Well take the first step to becoming a presenter by attending this session and I will guide you through the process of learning to present and becoming good at it. It's easier than you think! I am an introvert and was deathly afraid to speak in public. Now I love to present and it's actually my main function in my job at Microsoft. I'll share with you journey that lead me to speak at major conferences and the skills I learned along the way to become a good presenter and to get rid of the fear. You can do it!
General Session
Learning to Present and Becoming Good At It
James Serra
Microsoft
Session Goals: • Gain confidence to start presenting, even if you’re nervous or introverted. • Learn the key skills and steps to become an effective, engaging speaker. • Understand how presenting can accelerate your career and open new opportunities.
Session Prerequisites: • None.
Track: Professional Development
Level: Level 100
-
-
-
Room: TBD
Have you been thinking about presenting at a user group? Are you being asked to present at your work? Is learning to present one of the keys to advancing your career? Or do you just think it would be fun to present but you are too nervous to try it? Well take the first step to becoming a presenter by attending this session and I will guide you through the process of learning to present and becoming good at it. It's easier than you think! I am an introvert and was deathly afraid to speak in public. Now I love to present and it's actually my main function in my job at Microsoft. I'll share with you journey that lead me to speak at major conferences and the skills I learned along the way to become a good presenter and to get rid of the fear. You can do it! -
-
General Session
PostgreSQL Storage Internals: From MVCC to TOAST and Beyond
Adam Machanic
PostgreSQL has become a popular database choice thanks to its excellent programmability and permissive license. And while it tends to be easy to set up and manage, truly robust designs require an understanding of what's happening under the hood. In this session you'll learn about how Postgres handles one of the most important parts of any database engine, storing your data. The session begins with a discussion on Multi-Version Concurrency Control (MVCC). We will dive deep, examining how visibility rules, transaction snapshots, and hint bits are leveraged to enable high concurrency MVCC. Building on this, we'll explore tuple versioning in detail; how updates create new row versions, how dead tuples accumulate, and why vacuum is necessary to reclaim space. We'll also touch on write amplification, highlighting how your seemingly simple update can generate significant I/O overhead and storage bloat. Next the session will explore how Postgres handles complex data types–including arrays and composite structures–and how their representation influences both performance and space utilization. Finally, we'll cover TOAST (The Oversized-Attribute Storage Technique), explaining how large values are transparently managed, out-of-lined, and compressed, along with trade-offs this introduces. Attendees will leave with a mental model for understanding PostgreSQL's storage behavior and a better capability for both diagnosing performance issues and making informed design decisions.
General Session
PostgreSQL Storage Internals: From MVCC to TOAST and Beyond
Adam Machanic
Stealth Startup
Session Goals: • Understand the PostgreSQL storage system. • Understand how to make informed design decisions based on storage system tradeoffs.
Session Prerequisites: • Some knowledge of how database systems work (either SQL Server or Postgres). • Basic SQL.
Track: Database Management
Level: Level 300
Theme: Cloud + Multi-DB
-
-
-
Room: TBD
PostgreSQL has become a popular database choice thanks to its excellent programmability and permissive license. And while it tends to be easy to set up and manage, truly robust designs require an understanding of what's happening under the hood. In this session you'll learn about how Postgres handles one of the most important parts of any database engine, storing your data. The session begins with a discussion on Multi-Version Concurrency Control (MVCC). We will dive deep, examining how visibility rules, transaction snapshots, and hint bits are leveraged to enable high concurrency MVCC. Building on this, we'll explore tuple versioning in detail; how updates create new row versions, how dead tuples accumulate, and why vacuum is necessary to reclaim space. We'll also touch on write amplification, highlighting how your seemingly simple update can generate significant I/O overhead and storage bloat. Next the session will explore how Postgres handles complex data types–including arrays and composite structures–and how their representation influences both performance and space utilization. Finally, we'll cover TOAST (The Oversized-Attribute Storage Technique), explaining how large values are transparently managed, out-of-lined, and compressed, along with trade-offs this introduces. Attendees will leave with a mental model for understanding PostgreSQL's storage behavior and a better capability for both diagnosing performance issues and making informed design decisions. -
-
General Session
Airflow & dbt: The Hipsters of Data Integration Meet Azure
Tomislav Hlupić
Having been the indie darlings of the data space, Airflow and dbt have become mainstream – particularly in cloud environments where scalability and flexibility are king. They're revolutionizing the way we construct data pipelines with their fresh approach to orchestration and transformation, presenting a thrilling alternative to the legacy tools most of us are accustomed to. Now that Microsoft has acquired these new bright stars to include in its cloud ecosystem, there's an entirely new toolbox to tap into — without ever having to leave the confines of your existing stack. During this session, we will meet the dynamic duo of Airflow and dbt in Azure, see how they measure up to Microsoft Fabric pipelines, and review the pros and cons of each strategy. No matter if you're a seasoned coder or just exploring the cloud, you'll know a lot clearer where these tools play a role in your data adventure.
General Session
Airflow & dbt: The Hipsters of Data Integration Meet Azure
Tomislav Hlupić
Solvership
Session Goals: • Attendees will see where Airflow & dbt can be set up in Fabric. • How does dbt go versus more used approaches (Notebooks / Dataflows / SQL). • How can Airflow function in Fabric with the existing pipelines and notebooks.
Session Prerequisites: • Basic python knowledge just to understand the code.
Track: Development
Level: Level 200
Theme: Cloud + Multi-DB
-
-
-
Room: TBD
Having been the indie darlings of the data space, Airflow and dbt have become mainstream – particularly in cloud environments where scalability and flexibility are king. They're revolutionizing the way we construct data pipelines with their fresh approach to orchestration and transformation, presenting a thrilling alternative to the legacy tools most of us are accustomed to. Now that Microsoft has acquired these new bright stars to include in its cloud ecosystem, there's an entirely new toolbox to tap into — without ever having to leave the confines of your existing stack. During this session, we will meet the dynamic duo of Airflow and dbt in Azure, see how they measure up to Microsoft Fabric pipelines, and review the pros and cons of each strategy. No matter if you're a seasoned coder or just exploring the cloud, you'll know a lot clearer where these tools play a role in your data adventure. -
-
General Session
Architecting Sovereign Hybrid Postgres via Kubernetes-Native Design
David Wagoner
Modernization is no longer just about performance-it’s about sovereignty. As organizations adopt hybrid computing for AI, they face a "cloud trap": choosing between operational ease and proprietary vendor lock-in. This session introduces "Sovereign by Design," a Kubernetes-native approach ensuring your Postgres data layer remains portable across on-premises, private, and public clouds. We move beyond the traditional "DBA-as-a-person" model to a "DBA-as-an-Operator" paradigm using the CloudNativePG (CNPG) standard. Learn to architect "Five Nines" (99.999%) availability through control plane isolation and active-active Postgres Distributed (PGD) clusters. We dive into the mechanics of self-healing and automated Day 2 operations-high availability, scaling, and backups-without manual intervention. Finally, we discuss how a unified Hybrid Manager delivers a turnkey path to modernization, managing everything from single instances to global clusters via one interface. Build a database foundation owned by you, not your cloud provider. Learning Objectives: 1. Eliminate Lock-in: Master a sovereign architecture for database portability across any environment. 2. Automate Day 2: Implement the "DBA-as-an-Operator" model with CloudNativePG for automated resilience. 3. Architect Extreme Availability: Design active-active hybrid clusters for 99.999% uptime via control plane isolation.
General Session
Architecting Sovereign Hybrid Postgres via Kubernetes-Native Design
David Wagoner
EnterpriseDB
Session Goals: • Master sovereign architecture to eliminate vendor lock-in and ensure portability. • Implement the "DBA-as-an-Operator" model for automated resilience and Day 2 ops. • Design hybrid clusters via control plane isolation to achieve 99.999% availability.
Session Prerequisites: • Required: Understanding of Postgres administration and Kubernetes (pods, operators). • Recommended: Experience with enterprise or hybrid cloud to appreciate architectural trade-offs regarding sovereignty and resilience.
Track: Architecture
Level: Level 300
Theme: Cloud + Multi-DB
-
-
-
Room: TBD
Modernization is no longer just about performance-it’s about sovereignty. As organizations adopt hybrid computing for AI, they face a "cloud trap": choosing between operational ease and proprietary vendor lock-in. This session introduces "Sovereign by Design," a Kubernetes-native approach ensuring your Postgres data layer remains portable across on-premises, private, and public clouds. We move beyond the traditional "DBA-as-a-person" model to a "DBA-as-an-Operator" paradigm using the CloudNativePG (CNPG) standard. Learn to architect "Five Nines" (99.999%) availability through control plane isolation and active-active Postgres Distributed (PGD) clusters. We dive into the mechanics of self-healing and automated Day 2 operations-high availability, scaling, and backups-without manual intervention. Finally, we discuss how a unified Hybrid Manager delivers a turnkey path to modernization, managing everything from single instances to global clusters via one interface. Build a database foundation owned by you, not your cloud provider. Learning Objectives: 1. Eliminate Lock-in: Master a sovereign architecture for database portability across any environment. 2. Automate Day 2: Implement the "DBA-as-an-Operator" model with CloudNativePG for automated resilience. 3. Architect Extreme Availability: Design active-active hybrid clusters for 99.999% uptime via control plane isolation. -
-
General Session
Creating Agentic Solutions with Fabric IQ
Ginger Grant
AI isn’t getting a free pass for just answering questions anymore, it’s time to start putting it to work with Agentic AI. In Microsoft Fabric, this shift comes to life through Fabric IQ and Data Agents. Together, they form a practical, enterprise-ready foundation for completing tasks with your data and creating solutions instead of one-off AI experiments. In this session, you’ll see Fabric IQ woven directly into the platform, letting you interact with your data using natural language while getting real insights, not surface-level summaries. We’ll break down how it understands semantic models, metadata, and usage patterns so it can produce outputs you can actually use. Through live demos and repeatable design patterns, you see how to use different parts of Fabric to build an end-to-end agentic solution that connects insight, execution, and orchestration. We will also address the hard parts including security, role-based access control, and keeping agents within governed boundaries. You will walk out knowing how to build AI that doesn’t just answer questions, it gets things done.
General Session
Creating Agentic Solutions with Fabric IQ
Ginger Grant
Desert Isle Group
Session Goals: • Configure Fabric IQ to query semantic models using natural language and return structured outputs. • Build Data Agents that execute tasks against OneLake data using defined prompts and workflows. • Orchestrate multi-step agent workflows that combine Fabric IQ insights with Data Agent execution.
Session Prerequisites: • Knowledge of Fabric is helpful.
Track: Analytics
Level: Level 200
Theme: AI + Data
-
-
-
Room: TBD
AI isn’t getting a free pass for just answering questions anymore, it’s time to start putting it to work with Agentic AI. In Microsoft Fabric, this shift comes to life through Fabric IQ and Data Agents. Together, they form a practical, enterprise-ready foundation for completing tasks with your data and creating solutions instead of one-off AI experiments. In this session, you’ll see Fabric IQ woven directly into the platform, letting you interact with your data using natural language while getting real insights, not surface-level summaries. We’ll break down how it understands semantic models, metadata, and usage patterns so it can produce outputs you can actually use. Through live demos and repeatable design patterns, you see how to use different parts of Fabric to build an end-to-end agentic solution that connects insight, execution, and orchestration. We will also address the hard parts including security, role-based access control, and keeping agents within governed boundaries. You will walk out knowing how to build AI that doesn’t just answer questions, it gets things done. -
-
General Session
Building Secure Text-to-SQL Chatbots with PostgreSQL RLS
Durga Prasad
InduTeja Aligeti
Ankit Aryan
Can your AI chatbot accidentally leak sensitive data? As enterprises deploy LLM-powered Text-to-SQL interfaces, access control becomes critical. This session demonstrates how to build a chatbot that respects all data permissions through three defense layers. First, we implement schema pruning—dynamically filtering table and column metadata so the LLM never sees unauthorized structures. Second, we add AST-based validation using pg_query to parse and verify every generated SQL statement against allow-lists before execution, catching hallucinations and prompt injections. Finally, PostgreSQL's native Row-Level Security (RLS) provides the last line of defense, automatically filtering results based on user identity. Through live demos with a multi-tenant database, you'll see how a marketing user's query never returns HR payroll data, and how the system gracefully handles errors without leaking schema information. Walk away with a complete security blueprint for production AI chatbots.
General Session
Building Secure Text-to-SQL Chatbots with PostgreSQL RLS
Durga Prasad
Amazon
InduTeja Aligeti
AWS (Amazon Web Services)
Ankit Aryan
Amazon Web Services
Session Goals: • Architect a Text-to-SQL pipeline with role-based schema pruning and column-level access control. • Implement AST-based SQL validation to catch unauthorized queries before execution. • Configure PostgreSQL RLS policies that automatically enforce row-level security in chatbot responses.
Session Prerequisites: • Attendees should have experience with SQL databases and basic understanding of API development.
Track: Analytics
Level: Level 300
Theme: Security
-
-
-
Room: TBD
Can your AI chatbot accidentally leak sensitive data? As enterprises deploy LLM-powered Text-to-SQL interfaces, access control becomes critical. This session demonstrates how to build a chatbot that respects all data permissions through three defense layers. First, we implement schema pruning—dynamically filtering table and column metadata so the LLM never sees unauthorized structures. Second, we add AST-based validation using pg_query to parse and verify every generated SQL statement against allow-lists before execution, catching hallucinations and prompt injections. Finally, PostgreSQL's native Row-Level Security (RLS) provides the last line of defense, automatically filtering results based on user identity. Through live demos with a multi-tenant database, you'll see how a marketing user's query never returns HR payroll data, and how the system gracefully handles errors without leaking schema information. Walk away with a complete security blueprint for production AI chatbots. -
-
General Session
Observability for Microsoft Fabric: What's Happening in Your Platform
Jonathan Stewart
Which Fabric items consumed the most capacity units last week? When was your lakehouse table last refreshed? If a pipeline failed at 2 AM, would you know before your users told you? For most teams, the honest answer is: "I'd have to check three different places and hope I find it." Microsoft Fabric consolidates compute, storage, and analytics into one platform — but observability hasn't caught up. Capacity metrics live in one app, pipeline monitoring in another, semantic model health in a third. This session is for Fabric administrators, data engineers, and platform teams who need to answer "is my platform healthy?" without checking six dashboards. We'll build an observability practice from the ground up using Fabric's own APIs, XMLA endpoints, and capacity metrics: – Monitor capacity utilization and detect throttling before it impacts users – Track pipeline and notebook execution across workspaces with automated alerting on failures and SLA breaches – Audit semantic models continuously with the 78-rule Best Practice Analyzer — not once during development, but on every refresh – Measure data freshness across lakehouses and warehouses, flagging tables that miss their expected refresh windows This isn't a product demo. You'll leave with patterns, queries, and scripts built on the same building blocks available to every Fabric customer — ready to deploy Monday morning.
General Session
Observability for Microsoft Fabric: What's Happening in Your Platform
Jonathan Stewart
SoundBI
Session Goals: • Build a capacity monitoring practice using the Capacity Metrics App semantic model. • Implement automated pipeline and notebook failure detection across workspaces. • Connect capacity, pipeline, model health, and freshness signals into a unified platform health view.
Session Prerequisites: • Working knowledge of Microsoft Fabric administration or data engineering. • Familiarity with Fabric workspaces, pipelines, and semantic models. • Experience with capacity management or platform monitoring is helpful but not required.
Track: Analytics
Level: Level 300
Theme: AI + Data
-
-
-
Room: TBD
Which Fabric items consumed the most capacity units last week? When was your lakehouse table last refreshed? If a pipeline failed at 2 AM, would you know before your users told you? For most teams, the honest answer is: "I'd have to check three different places and hope I find it." Microsoft Fabric consolidates compute, storage, and analytics into one platform — but observability hasn't caught up. Capacity metrics live in one app, pipeline monitoring in another, semantic model health in a third. This session is for Fabric administrators, data engineers, and platform teams who need to answer "is my platform healthy?" without checking six dashboards. We'll build an observability practice from the ground up using Fabric's own APIs, XMLA endpoints, and capacity metrics: – Monitor capacity utilization and detect throttling before it impacts users – Track pipeline and notebook execution across workspaces with automated alerting on failures and SLA breaches – Audit semantic models continuously with the 78-rule Best Practice Analyzer — not once during development, but on every refresh – Measure data freshness across lakehouses and warehouses, flagging tables that miss their expected refresh windows This isn't a product demo. You'll leave with patterns, queries, and scripts built on the same building blocks available to every Fabric customer — ready to deploy Monday morning. -
-
Lightning Talk
7 Simple Strategies to Combat Imposter Syndrome
Amy Abel
Do you feel like a fraud? Do you constantly question your ability to do your job? This is imposter syndrome, and you are not alone. The reality is that most professionals in tech have navigated these feelings at some point in their careers. In this session, we will identify the warning signs of imposter syndrome and examine exactly how it can stall your career growth. I will share the specific strategies that have helped me overcome self-doubt throughout my own journey. You will leave with actionable steps to recognize your value, silence the inner critic, and own your success.
Lightning Talk
7 Simple Strategies to Combat Imposter Syndrome
Amy Abel
Straight Path
Session Goals: • Identify signs of imposter syndrome to recognize when self-doubt is clouding your judgment. • Understand how feelings of fraudulence can negatively impact your career growth and daily decision making. • Apply practical strategies and personal techniques to overcome the inner critic.
Session Prerequisites: • None.
Track: Professional Development
Level: Level 100
-
-
-
Room: TBD
Do you feel like a fraud? Do you constantly question your ability to do your job? This is imposter syndrome, and you are not alone. The reality is that most professionals in tech have navigated these feelings at some point in their careers. In this session, we will identify the warning signs of imposter syndrome and examine exactly how it can stall your career growth. I will share the specific strategies that have helped me overcome self-doubt throughout my own journey. You will leave with actionable steps to recognize your value, silence the inner critic, and own your success. -
-
General Session
Pokemon, Choose Your Index: You Can't Have Them All
Brent Ozar
Index tuning often sounds abstract: selectivity, key order, includes, trade-offs. In this session, let's turn it into a game! Delivered entirely through live demos and audience interaction, Brent Ozar turns index design into a series of Pokémon-style battles. Using the Users table from the Stack Overflow database, Brent Ozar presents real-world query patterns and invites the audience to “play” index cards—each representing a different index design—to see which ones win, which ones struggle, and which ones backfire. As each query battle unfolds, you’ll see why some indexes are super-effective against certain workloads but weak against others, why covering everything is tempting (and dangerous), and why you can’t catch—or create—every index without paying a price. Brent will explain the results in plain English, tying each outcome back to how SQL Server actually uses indexes under the covers. By the end of the session, you’ll walk away with a much stronger intuition for choosing the *right* index for a workload, understanding trade-offs, and explaining index decisions to developers and managers alike. No slides full of theory—just hands-on demos, real queries, and a room full of data professionals playing along.
General Session
Pokemon, Choose Your Index: You Can't Have Them All
Brent Ozar
Brent Ozar Unlimited
Session Goals: • Understand how index column order affects query processing. • Learn why you shouldn't start indexes with non-selective columns. • Discover how to craft indexes that are useful for all kinds of queries.
Session Prerequisites: • You should be comfortable reading T-SQL queries.
Track: Database Management
Level: Level 300
Theme: AI + Data
-
-
-
Room: TBD
Index tuning often sounds abstract: selectivity, key order, includes, trade-offs. In this session, let's turn it into a game! Delivered entirely through live demos and audience interaction, Brent Ozar turns index design into a series of Pokémon-style battles. Using the Users table from the Stack Overflow database, Brent Ozar presents real-world query patterns and invites the audience to “play” index cards—each representing a different index design—to see which ones win, which ones struggle, and which ones backfire. As each query battle unfolds, you’ll see why some indexes are super-effective against certain workloads but weak against others, why covering everything is tempting (and dangerous), and why you can’t catch—or create—every index without paying a price. Brent will explain the results in plain English, tying each outcome back to how SQL Server actually uses indexes under the covers. By the end of the session, you’ll walk away with a much stronger intuition for choosing the *right* index for a workload, understanding trade-offs, and explaining index decisions to developers and managers alike. No slides full of theory—just hands-on demos, real queries, and a room full of data professionals playing along. -
-
General Session
Defense in Depth: Securing Azure SQL Databases
Deepthi Goguri
Thinking about moving to the cloud but worried about data security? Securing your data requires more than just migration, it demands a strong, defense in depth strategy. In this demo driven session, you’ll learn how to protect your Azure SQL Database using practical, real world techniques. What this session will cover: 1. Network Security – Best practices for securing your network and managing access 2. Azure Role-Based Access Control (RBAC) – Assigning and managing user roles and permissions 3. Row-Level Security – Enforcing fine-grained access to sensitive data 4. Auditing – Monitoring database activity for compliance and accountability 5. Advanced Threat Protection – Detecting and responding to potential security threats in real time 6. Encryption Layers – Implementing Transparent Data Encryption (TDE) and Always Encrypted
General Session
Defense in Depth: Securing Azure SQL Databases
Deepthi Goguri
DBA
Session Goals: • Understand how to implement a layered, defense-in-depth security strategy for Azure SQL Database. • Learn practical techniques to protect data, control access, and monitor activity effectively. • Gain hands-on knowledge to confidently secure Azure SQL databases against threats and ensure compliance.
Session Prerequisites: • SQL Server and Azure SQL knowledge.
Track: Database Management
Level: Level 200
Theme: Security
-
-
-
Room: TBD
Thinking about moving to the cloud but worried about data security? Securing your data requires more than just migration, it demands a strong, defense in depth strategy. In this demo driven session, you’ll learn how to protect your Azure SQL Database using practical, real world techniques. What this session will cover: 1. Network Security – Best practices for securing your network and managing access 2. Azure Role-Based Access Control (RBAC) – Assigning and managing user roles and permissions 3. Row-Level Security – Enforcing fine-grained access to sensitive data 4. Auditing – Monitoring database activity for compliance and accountability 5. Advanced Threat Protection – Detecting and responding to potential security threats in real time 6. Encryption Layers – Implementing Transparent Data Encryption (TDE) and Always Encrypted -
-
General Session
SQL Server 2025 Vectors: Under the Hood and Into Production
Anthony Nocentino
SQL Server 2025's vector data type puts semantic search and retrieval-augmented generation directly into a data estate you already operate and trust. That's a big deal, and there's real engineering depth that is worth understanding before you ship anything to production. This session goes inside the vector data type: storage layout, index structures, and what actually happens at query execution time when you ask SQL Server to perform a similarity search. We'll be direct about where vector indexes help and where they don't. Then we'll zoom out and build a working example application that puts all of it together. If you're an architect, DBA, or developer building the next generation of SQL Server applications, this one's for you. Let's go.
General Session
SQL Server 2025 Vectors: Under the Hood and Into Production
Anthony Nocentino
Everpure
Session Goals: • Describe vector data type internals: storage, indexing, and query execution. • Evaluate vector index performance and limitations at scale. • Apply application architecture patterns to build a working application using SQL Server 2025's vector search capabilities.
Session Prerequisites: • Core database engine concepts like data types, indexes and reading query plans.
Track: Database Management
Level: Level 300
Theme: AI + Data
-
-
-
Room: TBD
SQL Server 2025's vector data type puts semantic search and retrieval-augmented generation directly into a data estate you already operate and trust. That's a big deal, and there's real engineering depth that is worth understanding before you ship anything to production. This session goes inside the vector data type: storage layout, index structures, and what actually happens at query execution time when you ask SQL Server to perform a similarity search. We'll be direct about where vector indexes help and where they don't. Then we'll zoom out and build a working example application that puts all of it together. If you're an architect, DBA, or developer building the next generation of SQL Server applications, this one's for you. Let's go. -
-
General Session
Read Committed Is Cruel by Default
Erik Darling
Nobody picked Read Committed. Ever. SQL Server came that way, and most shops never change it; they get scared off by blog posts that are somewhere between dumb and wrong, if they even know there's an alternative. And so it sits there locking everything in sight while you festoon every query with NOLOCK hints and hope for the best. Readers block writers, writers block readers, deadlocks never cease firing, and one report takes down the whole application. Row versioning has been in SQL Server since 2005: – Read Committed Snapshot Isolation flips the whole database over at once. – Snapshot Isolation lets you opt in one query at a time, so you can prove it works. If you think changing is risky, consider this: the moment Microsoft had to run customer workloads themselves, they switched RCSI on by default in Azure SQL Database. No isolation level is perfect for every single query. Your job is picking the one that causes the fewest problems for your workload, and handling the exceptions as they come up. If you already slap NOLOCK on everything in sight, using other locking hints to handle the exceptions shouldn't scare you a bit. Row versioning gives you lock-free reads without the dirt. This session shows you how to choose for yourself, instead of living with whatever the installer left you.
General Session
Read Committed Is Cruel by Default
Erik Darling
Darling Data
Session Goals: • Spot the blocking, deadlocks, and outright wrong query results that locking Read Committed produces by default, and why reaching for NOLOCK makes all of it worse. • Understand how Read Committed Snapshot Isolation and Snapshot Isolation fix that with row versioning, and what they cost you (the version store, update conflicts, and the few things they don't fix). • Leave able to choose row versioning for the bulk of a workload and handle the exceptions with targeted locking hints, instead of living with whatever the installer left on.
Session Prerequisites: • Comfortable reading and writing everyday T-SQL, plus a general sense of what a transaction is. • No prior knowledge of isolation levels, locking internals, or row versioning is needed; the session builds all of it up.
Track: Database Management
Level: Level 300
-
-
-
Room: TBD
Nobody picked Read Committed. Ever. SQL Server came that way, and most shops never change it; they get scared off by blog posts that are somewhere between dumb and wrong, if they even know there's an alternative. And so it sits there locking everything in sight while you festoon every query with NOLOCK hints and hope for the best. Readers block writers, writers block readers, deadlocks never cease firing, and one report takes down the whole application. Row versioning has been in SQL Server since 2005: – Read Committed Snapshot Isolation flips the whole database over at once. – Snapshot Isolation lets you opt in one query at a time, so you can prove it works. If you think changing is risky, consider this: the moment Microsoft had to run customer workloads themselves, they switched RCSI on by default in Azure SQL Database. No isolation level is perfect for every single query. Your job is picking the one that causes the fewest problems for your workload, and handling the exceptions as they come up. If you already slap NOLOCK on everything in sight, using other locking hints to handle the exceptions shouldn't scare you a bit. Row versioning gives you lock-free reads without the dirt. This session shows you how to choose for yourself, instead of living with whatever the installer left you. -
-
Lightning Talk
Synchronous Availability Groups Don’t Always Mean Zero Data Loss
Theodoros Katsimanis
Synchronous Availability Groups are often assumed to guarantee zero data loss. This session shows two real configurations where data can still be lost during crash/failover and how to configure AGs correctly when zero data loss truly matters.
Lightning Talk
Synchronous Availability Groups Don’t Always Mean Zero Data Loss
Theodoros Katsimanis
KaizenGaming
Session Goals: • Understand why synchronous commit does not always guarantee zero data loss. • Identify two configuration settings that allow committed data to be lost during failover. • Configure Availability Groups correctly when zero data loss is a strict requirement.
Session Prerequisites: • Basic understanding of SQL Server and Availability Groups is required. • Familiarity with transaction logs, commit behavior, and high availability concepts is helpful. • Some experience with production systems will make the session easier to follow.
Track: Database Management
Level: Level 300
Theme: Security
-
-
-
Room: TBD
Synchronous Availability Groups are often assumed to guarantee zero data loss. This session shows two real configurations where data can still be lost during crash/failover and how to configure AGs correctly when zero data loss truly matters. -
-
Deep Dive
An Inside Look at How Microsoft Builds SQL Features
Bob Ward
Come learn a behind the scenes look at how Microsoft builds features and capabilities for SQL. This session will go deep into how Microsoft uses AI technology from design to code to deployment including fun examples of how we debug our own product. This is an advanced session, but you will walk away with tips and techniques to apply to your use of SQL.
Deep Dive
An Inside Look at How Microsoft Builds SQL Features
Bob Ward
Microsoft
Session Goals: • Learn internals of important features of SQL Server. • Learn how Microsoft uses AI to build features to apply to your own application and system. • Have fun watching SQL experts debug and build the product on stage!
Session Prerequisites: • A fundamental understanding of SQL Server.
Track: Database Management
Level: Level 400
Theme: AI + Data
-
-
-
Room: TBD
Come learn a behind the scenes look at how Microsoft builds features and capabilities for SQL. This session will go deep into how Microsoft uses AI technology from design to code to deployment including fun examples of how we debug our own product. This is an advanced session, but you will walk away with tips and techniques to apply to your use of SQL. -
-
General Session
Exploring SQL Server 2025's Practical Benefits for T-SQL Developers
Andy Yun
2025 brought us a new release of SQL Server, but all you heard about was AI. Want to know if this release has any benefits for T-SQL developers? Then join me, where we'll explore the myriad of "meat and potatoes" offerings in SQL Server 2025. We'll cover topics that have a direct impact on T-SQL developers, like regular expressions, fuzzy string matching, and optimized sp_executesql. And we'll also explore enhancements that benefit performance, like optimized locking and feedback enhancements. When you leave, you'll have a clear appreciation for the practical enhancements in this release that benefit T-SQL developers. And hopefully agree with me that SQL Server just keeps getting better and better.
General Session
Exploring SQL Server 2025's Practical Benefits for T-SQL Developers
Andy Yun
Everpure
Session Goals: • Learn about the direct T-SQL enhancements that will directly impact a T-SQL developer, like RegEx and optimized sp_executesql. • Learn about the indirect enhancements in SQL Server 2025, that will improve one's T-SQL query performance, like optimized locking and IQP feedback enhancements.
Session Prerequisites: • Only general knowledge of SQL Server and T-SQL development.
Track: Development
Level: Level 200
Theme: AI + Data
-
-
-
Room: TBD
2025 brought us a new release of SQL Server, but all you heard about was AI. Want to know if this release has any benefits for T-SQL developers? Then join me, where we'll explore the myriad of "meat and potatoes" offerings in SQL Server 2025. We'll cover topics that have a direct impact on T-SQL developers, like regular expressions, fuzzy string matching, and optimized sp_executesql. And we'll also explore enhancements that benefit performance, like optimized locking and feedback enhancements. When you leave, you'll have a clear appreciation for the practical enhancements in this release that benefit T-SQL developers. And hopefully agree with me that SQL Server just keeps getting better and better. -
-
Deep Dive
Architecting "Agent-Ready" Data Platforms for Multi-DB Environments
Chanpreet Singh
The era of simple chatbots is evolving into the era of Agentic AI, but most enterprise data architectures remain "Human-Ready" rather than "Agent-Ready." High-functioning autonomous agents require more than just a vector database; they require a structured, observable, and governed foundation to prevent structural risk and operational failure. In this session, we move past the RAG to explore the architectural shift required to support autonomous agents across a Multi-DB ecosystem (SQL Server, PostgreSQL, and Cloud-Native stores). Drawing from real-world implementations and specialized research in automated monitoring for batch inference, we will deconstruct the "Engine Room" of modern AI. Attendees will learn: Agentic Data Engineering: How to build semantic metadata layers that allow agents to navigate complex schemas without human intervention. Foundations & Control: Implementing Observability frameworks to track agentic data retrieval, ensuring "Trust and Control" in high-volume environments. Structural Security: Applying Policy-as-Code and identity-based access to ensure agents adhere to strict governance and compliance standards. Operational Scaling: Strategies for managing the high-latency and cost challenges of running agents against hybrid-cloud data platforms. Whether you are managing traditional relational databases or modern cloud-native ecosystems, this session provides a vendor-neutral blueprint for bridging the gap between raw data and reliable AI agents.
Deep Dive
Architecting "Agent-Ready" Data Platforms for Multi-DB Environments
Chanpreet Singh
AWS
Session Goals: • Build Agent-Ready Foundations: Move from RAG to Agentic Data Engineering for Multi-DB navigation. • Automate Observability: Master frameworks for monitoring AI batch inference and trust. • Enforce Security: Apply Policy-as-Code to secure AI in hybrid-cloud architectures.
Session Prerequisites: • Attendees should have a foundational understanding of relational databases (SQL Server or PostgreSQL) and basic CI/CD concepts. • Familiarity with core AI terms like RAG or Vector Search is helpful, but not required.
Track: Architecture
Level: Level 300
Theme: Cloud + Multi-DB
-
-
-
Room: TBD
The era of simple chatbots is evolving into the era of Agentic AI, but most enterprise data architectures remain "Human-Ready" rather than "Agent-Ready." High-functioning autonomous agents require more than just a vector database; they require a structured, observable, and governed foundation to prevent structural risk and operational failure. In this session, we move past the RAG to explore the architectural shift required to support autonomous agents across a Multi-DB ecosystem (SQL Server, PostgreSQL, and Cloud-Native stores). Drawing from real-world implementations and specialized research in automated monitoring for batch inference, we will deconstruct the "Engine Room" of modern AI. Attendees will learn: Agentic Data Engineering: How to build semantic metadata layers that allow agents to navigate complex schemas without human intervention. Foundations & Control: Implementing Observability frameworks to track agentic data retrieval, ensuring "Trust and Control" in high-volume environments. Structural Security: Applying Policy-as-Code and identity-based access to ensure agents adhere to strict governance and compliance standards. Operational Scaling: Strategies for managing the high-latency and cost challenges of running agents against hybrid-cloud data platforms. Whether you are managing traditional relational databases or modern cloud-native ecosystems, this session provides a vendor-neutral blueprint for bridging the gap between raw data and reliable AI agents. -
-
General Session
Governing AI Agents: Designing Skills for Control and Trust
Tyler Garner
AI agents are rapidly being introduced into data platforms, but many implementations lack the governance, observability, and control required for enterprise use. Ad hoc prompts and loosely defined tool access create systems that are difficult to audit, unsafe to scale, and unreliable in production. This session introduces a structured approach to building “skills” as governed capability units for AI agents. Rather than allowing agents unrestricted access to databases, APIs, and workflows, we define skills with clear contracts, permissions, and operational boundaries. This enables teams to standardize how agents interact with data systems while maintaining control over execution and risk. We will walk through a reference architecture that separates agent reasoning from governed execution layers, showing how skills act as the interface between AI systems and enterprise data platforms. Attendees will see how to implement policy-driven access, enforce least-privilege patterns, and instrument workflows for observability and auditability. The session also covers practical considerations for production readiness, including logging, evaluation, failure handling, and feedback loops that improve system reliability over time. Examples will demonstrate how agents can safely query data, trigger workflows, and return results without introducing opaque or unmanageable behavior.
General Session
Governing AI Agents: Designing Skills for Control and Trust
Tyler Garner
Ion Storage Systems
Session Goals: • Understand how to design skills as governed interfaces for AI agents. • Learn to implement policy-driven access, observability, and auditability. • Apply patterns to transform experimental agents into controlled, production-ready systems.
Session Prerequisites: • Familiarity with SQL, basic API concepts, and general understanding of AI/LLM use cases. • No prior experience building AI agents is required, but attendees should be comfortable with architectural concepts and data system design patterns.
Track: Architecture
Level: Level 300
Theme: Security
-
-
-
Room: TBD
AI agents are rapidly being introduced into data platforms, but many implementations lack the governance, observability, and control required for enterprise use. Ad hoc prompts and loosely defined tool access create systems that are difficult to audit, unsafe to scale, and unreliable in production. This session introduces a structured approach to building “skills” as governed capability units for AI agents. Rather than allowing agents unrestricted access to databases, APIs, and workflows, we define skills with clear contracts, permissions, and operational boundaries. This enables teams to standardize how agents interact with data systems while maintaining control over execution and risk. We will walk through a reference architecture that separates agent reasoning from governed execution layers, showing how skills act as the interface between AI systems and enterprise data platforms. Attendees will see how to implement policy-driven access, enforce least-privilege patterns, and instrument workflows for observability and auditability. The session also covers practical considerations for production readiness, including logging, evaluation, failure handling, and feedback loops that improve system reliability over time. Examples will demonstrate how agents can safely query data, trigger workflows, and return results without introducing opaque or unmanageable behavior. -
-
General Session
Pick Your Poison: Choosing the Right Edition of SQL Server
Kevin Feasel
SQL Server 2025 reshapes the edition landscape. Standard Edition gains features that previously required Enterprise. Developer Standard Edition is a new option for dev/test environments. Web Edition is gone entirely. These changes affect licensing costs, architecture decisions, and migration planning. This session walks through every current SQL Server edition: what each one provides and where the limits are. We'll cover what changed in 2025 and how to choose the right edition for your workload. We'll also walk through the mechanics of upgrading or downgrading between editions, as well as licensing models. You'll leave with a decision framework for picking the right edition and a checklist for evaluating your current deployments.
General Session
Pick Your Poison: Choosing the Right Edition of SQL Server
Kevin Feasel
Catallaxy Services, LLC
Session Goals: • Learn what each edition of SQL Server includes and how they differ. • Learn how to move between editions of SQL Server, including on-premises and in Azure. • Understand when to choose which edition your own environment.
Session Prerequisites: • This session assumes some familiarity with SQL Server installation and features.
Track: Database Management
Level: Level 200
Theme: Security
-
-
-
Room: TBD
SQL Server 2025 reshapes the edition landscape. Standard Edition gains features that previously required Enterprise. Developer Standard Edition is a new option for dev/test environments. Web Edition is gone entirely. These changes affect licensing costs, architecture decisions, and migration planning. This session walks through every current SQL Server edition: what each one provides and where the limits are. We'll cover what changed in 2025 and how to choose the right edition for your workload. We'll also walk through the mechanics of upgrading or downgrading between editions, as well as licensing models. You'll leave with a decision framework for picking the right edition and a checklist for evaluating your current deployments. -
-
General Session
Inside Real SQL Server Breaches: Privilege Escalations and How to Stop It
Andreas Wolter
Fabiano Amorim
Real-world SQL Server attacks rarely follow textbook patterns – and many environments remain vulnerable to non-obvious privilege escalation paths that bypass typical security assumptions. In this session, we walk through realistic attack scenarios based on recently discovered SQL Server vulnerabilities. Starting from a low-privileged account, we demonstrate how attackers can escalate privileges to sysadmin, access sensitive data, and evade detection. Each scenario is immediately followed by a deep analysis of why the attack works in real environments – and how it could have been detected and prevented using native SQL Server capabilities. While some vulnerabilities demonstrated have been addressed or are considered by design, they expose broader patterns of misconfiguration and design assumptions that can lead to similar attack paths in modern environments. You will learn: • How privilege escalation paths in SQL Server actually work in practice • Why common security configurations and assumptions fail • How to detect suspicious behavior using auditing, T-SQL, and native tooling • What concrete steps you can take to harden your environment This session bridges the gap between offensive research and defensive practice – turning real attack techniques into actionable security improvements you can apply immediately.
General Session
Inside Real SQL Server Breaches: Privilege Escalations and How to Stop It
Andreas Wolter
Sarpedon Quality Lab LLC
Fabiano Amorim
Pythian
Session Goals: • How privilege escalation paths in SQL Server actually work in practice. • Why common security configurations and assumptions fail. • How to detect suspicious behavior using auditing, T-SQL, and native tooling. • What concrete steps you can take to harden your environment.
Session Prerequisites: • Basic understanding on SQL Server security features.
Track: Database Management
Level: Level 300
Theme: Security
-
-
-
Room: TBD
Real-world SQL Server attacks rarely follow textbook patterns – and many environments remain vulnerable to non-obvious privilege escalation paths that bypass typical security assumptions. In this session, we walk through realistic attack scenarios based on recently discovered SQL Server vulnerabilities. Starting from a low-privileged account, we demonstrate how attackers can escalate privileges to sysadmin, access sensitive data, and evade detection. Each scenario is immediately followed by a deep analysis of why the attack works in real environments – and how it could have been detected and prevented using native SQL Server capabilities. While some vulnerabilities demonstrated have been addressed or are considered by design, they expose broader patterns of misconfiguration and design assumptions that can lead to similar attack paths in modern environments. You will learn: • How privilege escalation paths in SQL Server actually work in practice • Why common security configurations and assumptions fail • How to detect suspicious behavior using auditing, T-SQL, and native tooling • What concrete steps you can take to harden your environment This session bridges the gap between offensive research and defensive practice – turning real attack techniques into actionable security improvements you can apply immediately. -
-
Learning Pathway
Adapting for AI Readiness as a Data Professional| No Fluff, Just Facts: AI Essentials for Data Professionals
Deborah Melkin
The AI hype train shows no signs of slowing down and it seems like every week there is a new model, a new acronym, or some new AI wizardry that commands our attention. We feel like we have to learn every new thing we hear about and it's difficult to sort fact from fiction when we become aware of these new things. Most importantly, how do we know what tools and technologies in the AI realm will help us and which ones are distractions? Spend an hour with an experienced data professional detailing what tools are helpful, what technologies are important to understand, and what are the essential AI skills to have as the job of a data professional evolves. No sales, no hype, and no fluff – just a session full of helpful information that you can apply immediately. We also hope you'll join us for sessions 2 and 3 of this learning pathway as offer more tangible guidance to navigating AI as a data professional.
Learning Pathway
Adapting for AI Readiness as a Data Professional| No Fluff, Just Facts: AI Essentials for Data Professionals
Deborah Melkin
GLG
Session Goals: • Learn which AI tools are helpful. • Learn which AI-related technologies and concepts are important to understand. • Learn what the essential AI skills are to have as you move forward in your career.
Session Prerequisites: • Basic knowledge of SQL is preferred.
Track: Analytics
Level: Level 200
Theme: AI + Data
-
-
-
Room: TBD
The AI hype train shows no signs of slowing down and it seems like every week there is a new model, a new acronym, or some new AI wizardry that commands our attention. We feel like we have to learn every new thing we hear about and it's difficult to sort fact from fiction when we become aware of these new things. Most importantly, how do we know what tools and technologies in the AI realm will help us and which ones are distractions? Spend an hour with an experienced data professional detailing what tools are helpful, what technologies are important to understand, and what are the essential AI skills to have as the job of a data professional evolves. No sales, no hype, and no fluff – just a session full of helpful information that you can apply immediately. We also hope you'll join us for sessions 2 and 3 of this learning pathway as offer more tangible guidance to navigating AI as a data professional. -