Saipraveen User

I’m an AI / Full-Stack Engineer with 4+ years of experience shipping LLM-integrated features into production. I focus on practical, measurable outcomes—building RAG systems, agentic workflows, and inference pipelines that perform reliably under real workloads.

In my recent role, I built RAG-based product search and customer support using LangChain, OpenAI embeddings, and Pinecone. I chunked and indexed catalogs and support docs, tuned semantic retrieval with MMR re-ranking, and reduced average support response time by 40%—while also reducing invalid AI-generated responses by 35%.

I’m hands-on with the “last mile” engineering around LLMs: structured output prompting, schema validation, fallback logic, and async inference optimization. I instrument token usage and prompt evaluation metrics to cut per-request token cost by 30%, and I improve perceived latency using streaming and caching (targeting under 800ms p95).

Beyond AI, I bring a strong full-stack and cloud foundation—TypeScript, Python, Node.js, React, and AWS microservices. I’ve deployed and monitored cloud-native systems with Docker and Kubernetes, automated CI/CD with GitHub Actions and Docker, and improved scalability and performance through caching, async processing, and database tuning.