Shilpa K

I’m an AI Engineer focused on building and deploying production-ready LLM applications, including Retrieval-Augmented Generation (RAG) and multi-agent systems. I care about reliability, grounding, and measurable outcomes—not just demos.

In my recent work, I fine-tuned a BERT-based classification model on 11,000+ records, improving accuracy from 78% to 87% and eliminating manual classification at load time. I also built FastAPI backends with optimized PostgreSQL queries to improve performance and deliver scalable REST APIs.

I’ve designed GPT-4-powered multi-agent workflows that automate business processes, including BRD generation with a ~90% reduction in turnaround time (from several hours to under 5 minutes). I apply prompt optimization and structured validation to keep outputs consistent and accurate.

On the project side, I’ve delivered end-to-end systems like an AI repository analysis platform (FastAPI + LangGraph + Gemini + PostgreSQL/pgvector) and a multi-source LLM-powered research tool (LangChain + ChromaDB + Gemini). I enjoy orchestration work, semantic search with embeddings, and containerized deployments that scale.