At KeelWorks Foundation, I’m planning and prototyping a retrieval-grounded Knowledge Assistant for users across nine disconnected internal knowledge systems. I audited the inherited LangGraph/FAISS prototype, catalogued seven defects, and established an allowlist, evaluation set, citation requirements, and a zero-cost deployment plan.
At Prodigy Infotech, I optimized PostgreSQL extraction across one million feedback records, improving dashboard query performance by 45%. I also exposed anomaly-detection inference through REST APIs and reduced image-classification preprocessing latency.
In my independent projects, I’ve built distributed services with FastAPI, Kafka, PostgreSQL, and Redis, and instrumented systems with OpenTelemetry and Prometheus. I’ve also contributed merged changes to SQLFluff, Arize Phoenix, and Great Expectations, including regression fixes and restored type checks.
