At Oracle, I built an AI incident investigation assistant that correlates error logs, request metrics, deployments, and code or configuration changes. Its workflow generates root-cause hypotheses and links diagnoses to supporting evidence.
I also connected diagnosis to recovery by selecting a rollback supported by deployment history and runbooks, requesting engineer approval, and checking service metrics afterward. At Navi Technologies, I developed an NL-to-SQL chatbot with a React and TypeScript interface and a LangChain/LangGraph workflow.
Earlier, as a Software Intern at Tofolo, I developed a PyTorch recommendation system and Kafka-backed feature pipelines. My projects include a GitHub issue-fixing agent that reproduces bugs with pytest and tests patches in Docker, and a customer support agent for order tracking, cancellations, and refunds.
