Databricks DevConnect is a technical meetup where data and AI practitioners dig into the details, collaborate, learn, and find their people, hosted by the Databricks Developer Relations team. Together, we’ll chat about new features, dive into meaty topics, share real-world experiences, and explore new ways to use Databricks so you can become the best developer you can be.
Why You Can't Miss This Event:
🏆 Become a 10x data and AI practitioner by using the best of Databricks for your Data + AI projects.
🧠 Learn from exclusive experts you won’t hear anywhere else, including Product Managers, DevRel, Field Engineers, and MVPs.
🤝 Connect with your community and build lasting relationships with fellow practitioners and Databricks Developer Advocates, MVPs, and Product teams.
👕 Get those goodies: Swag will be raffled off throughout the event and attendees will get access to hands-on training and guided labs through Databricks Academy Labs post event!
Session Descriptions
- Session #1: Lakebase - unlock operational use cases with agents and apps on your lakehouse data
- There are two types of database systems: OLAP for analytical use cases and OLTP for transactional operational systems. They go hand in hand. Apps and agents write to OLTP databases like Postgres, and later all company data is ingested into OLAP systems for analytics.But how do you power operational use cases that depend on analytical and machine learning data stored in the OLAP system? How do you sync data from OLAP back to OLTP for low-latency queries in apps and agents?Crossing system boundaries is always a governance and security headache, and custom pipelines are brittle. That’s where Lakebase comes in. It’s a new kind of OLTP database that deeply integrates with your data lakehouse.In this talk, we’ll demo Databricks’ new Lakebase and show how it enables operational use cases for AI agents and data apps — all within your Databricks workspace.
- Session #2: Simplify your stack so business users never message you again: Genie, Knowledge Assistant and Unity Catalog
- The "AI-ready" enterprise sounds great on a slide, but production reality often consists of disconnected chatbots, fragmented RAG implementations, and ungoverned data silos. To bridge this gap, you need a Compound AI system that leverages the full context of your data estate. Join us as we dive into how Unity Catalog serves as the unified governance layer and semantic store required to ground Databricks Genie for high-precision text-to-SQL tasks. We will demonstrate how to architect a Multi-Agent system using Agent Bricks, combining Genie with a Knowledge Assistant that autonomously navigates both your structured and unstructured data to provide real answers.
- Session #3: Operationalizing High Quality AI Agents: From Experimentation and Evaluation to Production with MLflow
- One of the top challenges in building a Gen AI agent is ensuring high-quality outputs and expecting the correct behavior in production. Building and experimenting with agents is easy; evaluating and putting them in production is hard. In this talk, we explore:
- The agent-driven development cycle and how to operationalize with MLflow Gen AI components
- Use MLflow scorers to evaluate and ensure high-quality agent results
- Demonstrate how MLflow GenAI components facilitate operationalizing agents for production
Venue