Tarun Gowda

About

Full-stack engineer who helped build and launch an internal genAI platform called GAIL, supporting multiple LLMs, confidential document upload for RAG pipelines, and collaborative chat. Worked across FastAPI, React/TypeScript, AWS/DynamoDB, and Azure, with notable ownership of backend RAG logic, MCP integration architecture, and frontend fixes that improved chat usability.

Key Strengths

• end-to-end ownership of internal genAI product backend with some frontend contributions
• built RAG pipeline infrastructure for confidential document workflows
• shipped first release in 3 months with a small 3-developer team
• strong full-stack problem solving in React and TypeScript
• designed resilient multi-stage file upload and indexing flow
• used custom hooks to simplify complex frontend state management
• sound backend architecture judgment for MCP integration
• security-minded API design with centralized key management and output sanitization
• disciplined engineering process with PR reviews, protected branches, CI, and QA
• effective production debugging approach using rollback-first mitigation
• improved chat UX by fixing non-dynamic autoscroll behavior across screen sizes
• Built and shipped a 0-1 GenAI product (EMULaiTOR) integrating PDF/text, qualitative, and quantitative data into an aggregated LLM output with supporting sources and charts
• Strong RAG + vector database implementation experience (Postgres + pgvector; top-k retrieval) and prompt engineering to synthesize results
• Pragmatic data quality approach for LLM systems (handling empty rows/columns; ensuring complete inputs to reduce hallucinations/unwanted answers)
• Production scalability/resilience mindset (region co-location to reduce latency, health checks, load balancer failover, horizontal scaling, async + autoscaling rules)
• Effective real-time debugging of agentic workflows via tracing, flags, structured API errors, and tool-call monitoring (identified wrong-tool selection issue and fixed logic/context)
• Able to lead and tailor technical workshops for mixed audiences (separate developer vs QA content) and lead a team despite being less experienced due to domain expertise
• Supports sales enablement and deal closing by translating technical differentiators for customer pitches (helped win a large pharma client)

Experience

Education

Skills

Languages