Built a multi-agent LLM system using LangChain with an orchestrator-memory-RAG architecture, delivering personalized career roadmaps to 50+ test users. → Full RAG pipeline: document ingestion → semantic chunking → embedding generation → vector retrieval → output formatting → Agentic memory adapting responses to real-time trends Stack: LangChain · OpenAI API · Anthropic API · RAG · Vector Retrieval