ROADMASS

Inspiration

Choosing courses for our concentration isn’t easy, especially with the busy lives we lead as college students. Often, we reach out for advice, only to receive generic responses that don’t really help us make informed choices. That’s where ROADMASS comes in. We created this tool to provide a clear, customized roadmap that acts like a mind map, guiding students to select courses based on their comments and needs. ROADMASS empowers students to make decisions with confidence and clarity.

What it does

ROADMASS acts as an AI chatbot that generates a graph based on user input. It can also create a graph for research resources that may be included in our concentration. For example, if we want to focus on machine learning, it will generate a graph of the best courses UMASS students should take and display a graph of relevant research resources for machine learning at UMASS.

How we built it

We use RAG to retrieve course content and research materials from UMASS websites. The FAISS library enhances course search functionality within RAG. LangChain and LangGraph are integrated to build the AI agent. Additionally, we use React and Next.js to publish our platform and create a chatbot interface.

Challenges we ran into

As freshmen tackling our first hackathon, we faced challenges due to limited experience in web development and AI model tuning. While building the site, we encountered persistent bugs that demanded considerable time and effort to resolve before everything functioned smoothly.

Accomplishments that we're proud of

We’re proud to have successfully completed our first project on time and made it fully functional. We’re also thrilled to have learned new skills, like Next.js and React, along the way.

What we learned

We learned how to use React, JavaScript, and Next.js, as well as how to build and debug a website. We also strengthened our persistence by working in a challenging environment for 36 continuous hours.

What's next for ROADMASS

We plan to revisit our code to improve the generation speed and performance. We'll also work on making the website more visually appealing and vibrant. In the future, we hope to expand our scope to a larger project, like creating course selection tools for multiple universities.