Data engineering insights, cost optimization guides, and tutorials from the LakeLogic team.
Five jobs: define, check, act, observe, remediate. Which tool fits each.
March 5, 2026 · 8 min readA runnable Kafka-to-Delta notebook: backlog throughput, event latency and rejected records, measured separately.
November 3, 2026 · on lakelogic.ioPolars expressions, Pandera, Patito, dataframely, GX and contracts, compared by what they do with bad rows.
March 5, 2026 · 7 min readA mesh splits the data, not the agreement. How to make the cross-domain interface checkable.
March 5, 2026 · 9 min readWhere serverless Spark, Polars, DuckDB, dbt and Ibis fit, and one definition of the rules across engines.
March 1, 2026 · 7 min readStop, drop or quarantine: how to choose per table, and the work quarantine still leaves you.
March 3, 2026 · 6 min readSchema tools check values too. The gap is what happens to a bad row or a new column.
March 3, 2026 · 8 min readWhat GX does well, how Expectations map to a contract, and when to stay.
March 3, 2026 · 12 min read