At Amazon, I architected a micro-batching ETL pipeline for ad server logs that cut ingestion latency from hourly to 10 minutes. I also designed a data lake architecture supporting a data mesh across domain-owned datasets.
I engineered AI agents that connect data lake systems for data discovery, reducing ad-hoc analytics effort by 50%. I led a team rebuilding the semantic and reporting layer, aligning more than 200 metric definitions and lineage.
At Deloitte, I built an anomaly detection data mart and dashboard for revenue leakage, accelerating month-end billing reconciliation by 80%. I also implemented Airflow orchestration for more than 100 daily jobs, improving pipeline reliability to 99.5%.
