At Google, I work on storage infrastructure for Procella, Napa, and Sawmill, supporting analytics, insights, and event data. I built the first version-skew validation framework to detect data-format regressions before release and help prevent severe outages.
I architected a canonical config registry that made configurations self-serve and led its adoption across four organizations and nine teams. I also led compression, compaction, and encoding optimizations that cut disk usage across 450+ tables by 28–43%.
Earlier, I architected Fi Money’s event streaming and observability platform and built core banking and onboarding microservices. At Amazon, I engineered ad-serving systems handling 10B+ daily requests within a 10 ms SLA.
