Cuny Hacks 2024: CUNYGPT!

Read more about the project on our devpost!

The chatbot was implemented using RAG architecture to provide contextually relevant answers from our vector database
We scraped ~17,000 courses being offered across all CUNY colleges in Fall 2024, as well as used a dataset of ~30,000 jobs and associated job skills
The AI rag chatbot was built using custom retrievers that interact with different vector databases and schemas to ensure it provides relevant responses by performing vector searches
Each message is embedded and the embedded message is used as a query vector to perform searches in respective vector databases
OpenAI, LangChain, and Vercel’s AI SDK were used to develop the backend

Timestamps:
0:00 - Team Introduction
0:50 - The Problem/Solution
1:18 - Frontend explanation
1:51 - Wireframe explanation
2:40 - Backend explanation
4:29 - Challenges we faced
5:01 - Landing/About page demo
8:15 - Chat Demo

Ignore tags:
#hackathon
#coding
#ai