Treating outputs as documented, queryable, reproducible products.

In-process analytics, fast ad-hoc SQL and file-native exploration.

Python services, clean APIs and production integrations.

GitHub APIs, automation and developer-facing tooling.

Practical LLM systems, tooling and automation in real teams.

Notes about the site itself and how it evolves.

Next.js app router, server components and edge-friendly rendering.

Columnar formats, storage layouts and efficient scans.

Glue code, services and data tooling that actually ships.

Vector retrieval for pragmatic RAG systems.

Retrieval-augmented generation grounded in real documentation.

Object storage patterns, pushdown and large-scale data access.

Query design, optimization and SQL as an interface to data.