RAG Works Better When Your Data Is Already Tidy
A practical take on retrieval pipelines: tidy documents, useful metadata, and fewer magic promises.
Read the articleBlog
Future-proof thinking for data teams.
Short, useful essays on systems design, data platforms, and applied AI. Built for senior engineers who value clarity.
Dagster Pipelines Need Clear Ownership
A short note on why orchestration works better when assets, alerts, and ownership are obvious.
Read the articleBronze to Silver Is Where Trust Starts
Why the bronze-to-silver step matters more than people think when building reliable data products.
Read the articleIceberg and Nessie Make Data Changes Safer
Why versioned tables and catalog branches make data changes easier to test and safer to release.
Read the articleData Engineering Best Practices
Blueprints for reliability, observability, and evolving pipelines with confidence.
Read the articlePython Logging That Actually Helps
Patterns to reduce noise and produce logs that tell a story during incidents.
Read the articleWriting Themes
What you can expect over the next months.
Reliable Pipelines
Design patterns, testing strategies, and operational checklists.
Platform Thinking
Developer experience, internal tooling, and long-term maintainability.
AI in Production
How to make AI useful without inflating your architecture.