AI Agent / RAG 应用原型系统，基于 Python + Streamlit + ChromaDB 构建，支持多轮对话、文档问答、向量检索、工具调用、模型切换与 Docker 部署。

RAG-based mutual fund analysis tool for Indian investors. Fetches real AMFI data, computes CAGR · Sharpe · Sortino, and answers natural language queries with structured fund comparisons using LLaMA 3.3 + ChromaDB + Streamlit. Dockerised multi-container setup, Kubernetes manifests, and GitHub Actions CI/CD pipeline.

RAG-powered AI chatbot for spiritual guidance using personal Obsidian notes. Built with FastAPI, Next.js 14, and vector embeddings for semantic search and context-aware conversations.

AI-assisted regression triage platform with semantic retrieval, vector similarity search, and Ollama-powered reasoning workflows.

RAG Q&A + autonomous triage agent for customer feedback. Built on Gemini free tier with ChromaDB, structured JSON validation, and deterministic routing.

🦜🔗 Self-correcting Agentic RAG agent built with LangGraph — combines Adaptive RAG, Corrective RAG (CRAG) & Self-RAG. Routes queries to vector store or web search, grades document relevance, and detects hallucinations before answering. Powered by Gemini 2.5 Flash, ChromaDB & Tavily.