At JellyGit, I develop AI-powered backend systems and scalable APIs using Python, TypeScript, and FastAPI. I also build LLM-powered chatbots and RAG applications for semantic search and document understanding.
I develop AI agent and multi-agent workflows with LangChain and LangGraph, and integrate OpenAI APIs into enterprise applications and automated business workflows.
In helperBot, I built a LangGraph customer support agent with a five-stage workflow, Chroma vector search, and cross-encoder reranking. I added a human approval step for refund workflows and an evaluation pipeline for response quality and safety.
I also built and deployed CloudOptimizer, a multi-tenant SaaS dashboard that analyzes AWS costs and sends alerts. I implemented per-user AWS credentials and fixed missing authentication on three routes so each dashboard displays only that user's data.
