At AITorch Consultants, I develop autonomous AI agents using OpenAI Agents SDK, LangGraph, and Model Context Protocol (MCP) to automate multi-step workflows and decision-making tasks.
I also design multi-agent orchestration workflows in LangGraph, implementing memory management, state transitions, task planning, and human-in-the-loop interactions.
I developed an LLM-based insurance claim validation system using LangChain, FAISS, and FastAPI to check claims against 1,000+ policy clauses. Its semantic search and vector retrieval pipelines reduced policy lookup time by 70% with sub-two-second response latency.
I built a RAG customer support assistant with sentence-transformer embeddings and GPT-3.5-Turbo. It retrieves relevant knowledge-base documents to provide context-aware answers and reduce hallucinations.
