Field notes
On nested data, drift, and deciding what's actually worth asserting. Written to be useful whether or not you ever use this tool.
- Data engineering
Flattening nested JSON for SQL: the three decisions that matter
Grain, empty arrays, and name collisions — the choices that get made implicitly the first time and rewritten later. With DuckDB, Snowflake, and BigQuery syntax.
Read post → - Data engineering
Detecting schema drift before it breaks your pipeline
The drift that throws errors is the safe kind. It's the silent presence and distribution shifts that corrupt reports for weeks before anyone notices.
Read post → - Webhook ingestion
Why webhook payloads break pipelines, and what to check
Per-event shapes, genuinely optional fields, at-least-once delivery, and vendor changes nobody announces — plus which checks belong at the door and which belong after landing.
Read post → - Analytics engineer
Which dbt tests are actually worth writing
Most suites test every primary key and nothing else. Why relationships is the highest-value test you're skipping, and why range tests cause most self-inflicted alert fatigue.
Read post →