I've resolved 30+ production issues in Oracle MySQL-InnoDB across transactions, recovery, indexing, concurrency, full-text search, and performance. I also prototyped multi-page undo logging that increased large-update transaction capacity by approximately 16× while preserving rollback and crash recovery correctness.
At Uber For Business, I analyzed ride, ticket, and operational data in SQL to identify a validation defect that helped prevent approximately $220K in recurring monthly leakage. I also maintained 10–12 production ETL pipelines and scheduled jobs.
I've built systems tooling including Rediscope, a Redis RDB byte-level explorer with 180+ npm downloads, and a distributed key-value store that handled 100K writes with zero data loss during single-node failures.
I bring hands-on experience debugging Linux and database internals, reasoning about failure modes, and building observable backend systems with Docker, Prometheus, and Grafana.
