Hudson is an attempt to build an AI interface version of myself. An OpenAI GPT-4 RAG based chatbot that can answer questions about my CV, blog posts, and more.
The front-end of Hudson is a Next.js web app paired with the Vercel AI SDK, providing a chat UI to interact with the AI.
The back-end is Node.js app pairing LangChain with a Supabase Postgres vector database to implement Retriveval Augmented Generation (RAG) on top of the Open AI GPT-4 model.
The initial use case is to provide an AI version of my CV that you can talk to. Hudson has access to my CV and details on the specifics of what I achieved in each role. It also has access to a corpus of my thoughts & views from blog posts and specific markdown files which have been provided during RAG.