How Hamlet uses Databricks Lakebase Postgres to test changes against production data

Hamlet, an AI platform indexing local government meetings, relies on Databricks Lakebase Postgres to create isolated data branches for each engineering pull request. This allows developers and AI agents to test changes against real production data without altering the live environment. Each branch remains connected to a preview deployment, enabling immediate validation of code changes. A human reviewer still approves every merge, maintaining oversight.

The company aggregates and standardizes meeting data from over 7,200 governing bodies across the U.S., including city councils, planning commissions, and zoning boards. Its platform helps users search transcripts, track policy developments, and set alerts for specific topics or addresses. Hamlet’s CEO Sunil Rajaraman notes the fragmented nature of local government websites complicates data collection, requiring integration with multiple CMS providers and handling diverse formats like PDFs and scanned documents.

Hamlet’s customers, primarily real estate and data center developers, use the platform to identify project approval patterns and precedents. The system identifies speakers, timestamps key discussions, and provides structured insights to help users understand officials’ positions. AI models analyze sentiment and context to determine support or opposition for specific issues, such as housing or data centers, with multiple reviews ensuring accuracy.

Hamlet previously used AWS RDS with separate production and staging instances, which required 15 to 30 minutes of manual setup per pull request to align data for testing. Lead Developer Erich Douglass highlights that Lakebase’s data branching feature eliminated this overhead, allowing the small team to focus on velocity. Engineers now test against production-like data directly, reducing delays and improving the reliability of their pipelines.