/pitch
Where the work happens — selected projects and a short summary of what each one solved.
Ingestion and transformation pipeline pulling from multiple sources into PostgreSQL. Handles gap-filling and validation so downstream dashboards never show stale or broken data, and runs serverless to keep costs near zero when idle.
Multi-source reporting built on Apache Superset, backed by carefully tuned SQL with correct per-entity scoping so each view answers exactly one question without double-counting.
A serverless gateway sitting in front of a slow, rate-limited upstream API. Uses a cache with stale-while-revalidate plus single-flight locking to stay fast and stay within limits under load.