PV PathFinder

The shortage of advisors at PVAMU makes it a hassle to know your degree. However, with our AI-powered advising app, PV PathFinder, students can take a hold of of their education right at home.

Comment

Inspiration

We were inspired because it's course registration season, and getting access to advisors is always a hassle. The singular computer science advisor at PVAMU left, and so we've been having to share an advisor with engineers, overworking them.

What it does

Out product allows a student to input their major, minor, courses they've taken, and when they want to graduate, and gives a student a detailed degree plan with specific courses to take each semester, as well as an explaination. This product also allows the user to note any special circumstances that may change their degree plan, which allows for great flexibility. Each user gets a plan completely unique to them.

How we built it

We built this by creating a website supported by a flask backend. We send the user's input to flask, which runs it through google gemini, and then gives the user gemini's generated plan. In order for this to be possible, we compiled a list of every course offered at PV, every major's degree plan, and every minor's requirements. This pdf was initially over 1000 pages long, but we shrunk it down to under 100.

Challenges we ran into

The amount of data we had to give the LLM to make this possible was a struggle. Many times, the LLM said our request was too large, so we had to shorten our text-based data into the most concise form possible.

Accomplishments that we're proud of

We were stuck on an issue with Python's reading of the pdf file, but we were able to back track and find the root cause of the issue. We were able to gain an understanding of how the API collected and recieved data, and we were able to get flask running fairly quickly, even though this was our 2nd time using it . This is also only our 2nd hackathon, and I'm proud that we were able to finish out project without being stressed about not getting it in on time.

What we learned

We learned a lot about breaking down problems into small tasks, as I feel like that helped us progress more smoothly. We also learned a lot about how web apps run and how to use APIs in our projects for the future.

What's next for PV PathFinder

We're thinking about pitching this idea to PV and seeing if they could make it a reality, as we believe it would truly benefit the students. While every student has access to AI, there aren't any AI that have PV's entire academic catalog at their disposal. Thus would make advising more convenient for both the students and the advisors.

Share this project:

Updates