In this project we learned how to use Retrieval-Augmented Generation (RAG) to improve AI-powered applications to provide accurate, context-aware responses. Built using Next.js, Clerk for auth, Pinecone, and the OpenAI API.
- Vector Database (Pinecone): We store and manage professor reviews as vectors which allows efficient search and retrieval.
- OpenAI for Language Understanding: OpenAI powers the natural language responses so that it makes the interactions feel more human-like.
- Mock Data: We use sample reviews to simulate real-world usage to demonstrate how RAG can imporve decision-making by providing detailed insights.
The following is an example of an interaction between a chatbot and a user
test