Hi, I'm a backend-focused developer specializing in Python, FastAPI, and production-grade AI systems - including multi-agent orchestration, RAG pipelines, and LLM integration. One of my recent projects was an AI-powered multi-agent medical assistant for healthcare professionals, researchers, and patients. In a clinical context, a confident-but-wrong answer is dangerous, and most chatbots will happily hallucinate a diagnosis. So I built a system where specialized AI agents answer medical questions safely - with sources and a human in the loop. Core features and tech: - Intent- and confidence-based routing to the right specialist agent - A RAG agent pulling answers from medical literature via semantic search and reranking - Vision agents analyzing X-rays, brain tumors, and skin lesions - Real-time clinical research retrieval - Mandatory expert verification before any recommendation Built with Python, FastAPI, LangGraph, Qdrant, Azure OpenAI (GPT-4o), and Docker. The hardest part was preventing hallucinations. A single model answering everything breaks the moment a question falls outside its training. I solved this with graph-based agentic workflows plus confidence scoring and guardrails - so the system knows when it doesn't know, defers to the right specialist, cites its sources, and never bypasses human review. I added hybrid search and cross-encoder reranking to keep retrieved context genuinely relevant, which sharply cut wrong answers. My strength is combining scalable backend architecture with safe, practical AI integration - building systems that are not just smart, but reliable, observable, and ready for real production use in demanding domains. If you're interested in my skills and have a project or idea in mind, feel free to reach out - I'd be glad to help.