Loading…
Ask, build, compose: What our 5th Genie Hackathon taught us about Databricks Genie
Shruti Prasanna, Rob Bajra
- Source
- Databricks
- Published
- Added to Yomu
Summary
The fifth Databricks hackathon used Databricks Genie to test three ways of working with governed data: asking, building, and composing. Genie Agents give business users domain-specific natural-language access to curated data, while Genie Code helps analysts create metric views, functions, pipelines, and dashboards inside Databricks. The composition track uses Conversation APIs and a managed MCP server to make Genie an agent tool alongside custom logic, Model Serving, and other MCP servers, with Unity Catalog governing access. Projects included supervisory routing across 190 tables, governance and analytics products built in days, and multi-agent systems such as ShipBob’s overnight operations brief. Across the tracks, the stated lesson is that shared governance and semantic context let business users, builders, and engineers use Genie at different levels without abandoning grounded, permissioned data access.
Context
The hackathon was run because building is described as the fastest way to learn a product. This edition focused on Databricks Genie, whose family serves business users, analysts and builders, and engineers composing agents.
Approach / What changed
The event organized projects into three tracks: Genie Agents for asking questions of governed data, Genie Code for creating analytics and data assets, and Genie Agents composed as tools within Databricks Apps and broader agent systems. Unity Catalog governance and Genie Ontology provided shared control and semantic understanding across the tracks.
Takeaways
- A single Genie Agent is designed for up to 30 tables; OneTrust handled 190 tables and more than 300 views by sharding data across focused agents, routing questions, and stitching answers together.
- Genie Code supported projects that built metric views, Unity Catalog functions, pipelines, dashboards, and AI-assisted analytics inside Databricks without a separate development environment.
- Genie Conversation APIs and Databricks’ managed MCP server let agents call Genie as a governed tool alongside custom logic, Model Serving endpoints, and other MCP servers, with OAuth, Unity Catalog, and MLflow used for control or tracing.