Peter Fan

my journey exploring the world of information technology

Projects

LLM Agent Chatbot

Since the my last time building a chatbot, the field of Agentic AI has advanced significantly. In this project, I built a chatbot using a multi-step, reasoning LLM that can use various tools (ie. RAG tool, math tool, etc.) to better answer the user's question. The chatbot can also remember past interactions, making it more conversational and personalized.

Tech Stack
LangGraphOpenAIGoogle GeminiReAct AgentPinecone
Tags
AIChatbotLLM AgentRAGTool Calling

GPT-3 Chatbot (deprecated)

chatbot_screenshot
🤖 This version is deprecated

I engineered a RAG LLM chatbot that can answer questions about myself. Through this hobby project, I learned how to use OpenAI's embedding model and the Langchain framework to create a vectorized knowledge base (KB), host the vectorized knowledge on a vector database, and query the database using a Conversational Q&A Chain. It was a fun project that taught me a lot about the challenges in prompt engineering and the importance of a good knowledge base using structured data.

Tech Stack
LangchainOpenAIPythonReactPinecone
Tags
AIChatbotOpenAILLMPrompt EngineeringRAG

NextJS Website

Instead of using a cut-and-paste template, I decided to build my website from scratch using NextJS and Tailwind CSS. I learned a lot about server-side versus client-side rendering, edge computing, and different web app hosting solutions.

Tech Stack
ReactTypeScriptNext.jsTailwind CSSFirebase
Tags
Web DevelopmentRESTful API