Vishnu Tallapar eddy

Architected and shipped end-to-end AI-powered medical chatbot with multiple personas (patient assistant, clinician support, administrative bot) using LangChain and LangGraph, processing 100K+ conversations monthly with high accuracy. Built production RAG pipeline with 10M+ documents achieving strong retrieval precision and sub-2s latency. Implemented agentic workflows, security controls, and HIPAA-compliant data handling; led architecture reviews and mentored a team of engineers. Delivered a responsive frontend with real-time streaming and multi-turn context management.

Architected and shipped multi-persona medical chatbot system using LangChain and LangGraph with GPT-4 and Pinecone; built end-to-end backend on Azure (FastAPI, PostgreSQL, Redis) with HIPAA-compliant data handling and audit logging; implemented real-time frontend streaming with Next.js; led a team of 4 engineers and delivered features with 94% accuracy validated by medical experts.

Built an enterprise LLM platform and agentic systems delivering 99.9% availability for 500K+ users. Implemented multi-agent coordination (analysis, decision, execution) with tool-calling, semantic caching, and real-time document processing. Reduced API costs by optimizing embeddings and query paths; developed REST/GraphQL services and an evaluation framework with LangSmith for ongoing quality and cost insights.