With Course-Pilot students are able to instantly query their course materials using a RAG-based (Retrieval-Augmented Generation) AI assistant. This project was developed as part of the University of Calgary's AI Research School program in a team of 3 members.
My Role
The following contributions were made by me in the early iteration of the project:
Designed and implemented the all of the user interface, including the main page and chatbox page.
Created a dynamic dashboard that includes buttons for each course chatbot which is updated as courses are created.
Implemented a delete function that allows users to delete course chatbots.
Assisted with the front-end and back-end integration.
Presented project updates and technical details to research supervisor.
Key Features
Frontend (Next.js 15)
Course Dashboard: A clean, responsive grid layout for managing multiple courses.
Course Management: Easily add new courses or delete old ones directly from the UI.
Backend (FastAPI + ChromaDB)
Local RAG Pipeline: Runs entirely on your machine using Ollama (Mistral/Llama 3) — no OpenAI API keys required.
Robust Retrieval: Normalizes Course IDs (e.g., handles "CPSC 383" vs "CPSC383") ensuring queries always find their target data.