Built backend services at TestMu AI (LambdaTest) for agent-to-agent testing across chat, voice and phone agents using Python, FastAPI, PostgreSQL and Kafka.
Designed REST APIs and background test-run tracking so users could follow results live. Improved processing under duplicate updates and timeouts to keep run status and results consistent.
Worked on LLM evaluation settings and prompts, Slack and email alerts, organization-scoped agent context, evaluation data and LLM usage billing.