ColRAG

Inspiration

As students, we often struggle to find quality content on platforms like Chegg and Quizlet. These platforms only provide answers to existing questions but fail to prepare us for professor-specific or class-specific exams, tests, and overall class readiness. This gap in personalized study resources inspired us to create an application that addresses this issue using a Generative AI RAG (Retrieval-Augmented Generation) chatbot.

What it does

Our Generative AI RAG chatbot helps students prepare for professor-specific and class-specific exams by providing personalized study resources, unlike platforms like Chegg and Quizlet that only offer generic answers.

How we built it

We developed this project using:

MERN Stack (MongoDB, Express.js, React, Node.js) OpenAI API for generating responses MongoDB Atlas for cloud database management Mongoose library for database modeling LangChain for enhanced language model integration Multiple AI models to improve accuracy and reliability

Challenges we ran into

Purchasing OpenAI API Tokens – Managing API costs while ensuring optimal performance. Setting Up MongoDB Atlas – Integrating a scalable cloud database into our backend. Generating YouTube Links for Class Topics – Ensuring OpenAI retrieves relevant educational videos to enhance learning support. By overcoming these challenges, we built an AI-powered study tool that enhances student learning beyond generic question-answer platforms.

Accomplishments that we're proud of

Customized Learning – Developed a chatbot that tailors study materials to specific professors and classes.

Advanced AI Integration – Successfully implemented Retrieval-Augmented Generation (RAG) to provide accurate and context-aware responses.

Bridging the Study Gap – Addressed a real student pain point that existing platforms like Chegg and Quizlet fail to solve.

Efficient Data Processing – Designed a system that effectively reads and understands parsed text files.

User-Centered Design – Created an intuitive and accessible interface for students to easily interact with the AI.

Scalability & Adaptability – Built a framework that can expand to support multiple courses and institutions.

What we learned

Retrieval-Augmented Generation (RAG) in Practice – Learned how to implement RAG to enhance chatbot accuracy and relevance.

Data Processing & Parsing – Gained experience in handling and structuring parsed text files for efficient AI retrieval.

Challenges of AI Accuracy – Ensured responses remain relevant and factual, avoiding AI hallucinations.

Collaborative Problem-Solving – Working as a team helped refine ideas and improve the chatbot’s effectiveness.

What's next for ColRAG

We plan to work with our university to make ColRAG a valuable study resource by personalizing the app according to the classes offered and the professors teaching those courses. This will help improve the quality of education by providing tailored study materials. We also aim to collaborate with our university's Innovation Hub to secure potential funding, which will allow us to make modifications to the app and scale it further.