How we built Cloudflare's data platform and an AI agent on top of it

This post details Cloudflare’s Town Lake unified data lakehouse and Skipper, an AI-driven data agent: their architecture (Trino query engine, R2/Iceberg storage, DataHub metadata, Lifeguard access control, Skimmer PII scanner, Transformer ELT, ingestion pipelines), default-closed governance, and how Skipper translates natural-language queries into audited SQL and visualizations while enforcing per-session PII redaction and user-based access controls to support billing, BI, security analytics, and self-serve data engineering.