Web app to ask AI questions about my college course material.
https://ask-ai.varunbalani.com/
This project performs Retrieval Augmented Generation(RAG) using embeddings generated from my course material. Embeddings were generated using OpenAI's text embedding model: text-embedding-ada-002.
The generated vectors are stored in a Pinecone index with different namespaces corresponding to each subject. When a user enters a query, a similarity search is performed on the stored embeddings
for the selected subjects. The top-k results are returned, and given as context to the chat completion model which uses the context to answer the user's query.
- React
- Express
- PostgreSQL with Prisma
- Pinecone as a Vector Database
- Wasp-Lang(https://wasp-lang.dev/)
