A production-ready RAG-powered AI chatbot embedded directly into my portfolio, allowing visitors to ask questions about my skills, experience, projects and contact information through a conversational interface.
I built a FastAPI-based AI assistant connected directly to the portfolio frontend. Each visitor receives an isolated session ID, messages are sent through the /chat endpoint, conversation context is managed through a three-phase hybrid memory system, and LangChain connects the structured portfolio context with Gemini to generate relevant responses. Strict system instructions keep answers focused on my real portfolio information and reduce hallucination.
Click Play a Run to follow the system from start to finish.
Portfolio AI Assistant is a conversational AI chatbot I built for my portfolio website using FastAPI, LangChain and Google Gemini. Visitors can ask questions about my technical skills, professional experience, projects and contact details without manually navigating through the entire website. The assistant maintains isolated conversation sessions and uses a hybrid summarization memory system to preserve context while keeping prompts compact.
A traditional portfolio requires recruiters and clients to manually browse different sections to find specific information. Visitors may want quick answers about technologies, experience, individual projects or how to contact me without searching through multiple pages.
Embedded portfolio chat assistant
FastAPI REST API backend
POST /chat conversational endpoint
Health-check API endpoint
Google Gemini response generation
LangChain LLM integration
Portfolio-aware system prompt
Answers about skills and technologies
Answers about professional experience
Answers about portfolio projects
Answers about contact information
Unique session ID for each visitor
Session-isolated conversations
Three-phase hybrid conversation memory
Progressive conversation summarization
Maximum five stored summaries per session
Reduced prompt-context growth
Strict portfolio-only response rules
Input validation
Empty-message handling
Long-message validation
Graceful LLM failure handling
CORS support for Next.js integration
Structured application logging
Floating chatbot interface
Render backend deployment
Next.js portfolio integration
I built a FastAPI-based AI assistant connected directly to the portfolio frontend. Each visitor receives an isolated session ID, messages are sent through the /chat endpoint, conversation context is managed through a three-phase hybrid memory system, and LangChain connects the structured portfolio context with Gemini to generate relevant responses. Strict system instructions keep answers focused on my real portfolio information and reduce hallucination.
