Built an interactive portfolio featuring HiteshBot, a custom AI assistant trained on my professional background. Uses RAG and LLMs to provide context-aware responses in real-time.
This portfolio showcases a modern application of Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG). HiteshBot is a specialized AI assistant that can answer complex questions about my professional background, projects, and experience with high accuracy.
The backend is built with FastAPI and LlamaIndex, using Google's Gemini Pro for both conversational responses and vector embeddings. The system features real-time streaming, intent recognition, and strictly grounded responses to ensure professional and accurate interactions.
Key features include conversational AI, RAG-based knowledge retrieval, advanced prompt engineering, and a custom data pipeline for ingesting structured resume information. The frontend is built with Next.js and TypeScript, providing a modern and responsive user interface.