Ajith P

About

Backend/AI engineer with healthcare domain experience who built a patient journey analytics API (FastAPI/PostgreSQL/Snowflake/Redis) and debugged peak-hour latency down from ~900ms to ~50ms via indexing and query optimization. Shipped an LLM-powered clinical summary/recommendation assistant end-to-end and designed a multi-step risk evaluation agent workflow with safety guardrails against hallucinations and unsafe outputs.

Experience

Education

Key Strengths

• Diagnosed and resolved peak-hour API latency via observability + DB tuning (P95 tied to DB; query reduced ~900ms to ~50ms with composite index)
• End-to-end ownership of LLM-powered clinical summary feature (stakeholder discovery through production monitoring)
• Strong data-layer performance optimization (index scans, limiting selected columns, pagination)
• Designs reliable multi-step agent/workflow systems with planner vs execution separation
• Robust error handling strategy (transient vs recoverable vs critical; exponential backoff retries; safe fallbacks for unsafe AI outputs)
• Production-grade Python practices (clean architecture, typing/Pydantic validation, async concurrency, testing, structured logging with correlation IDs)