AI Advisor

Our Inspiration!

The inspiration is that, as a student, we have a lot of frustrations with the process of schedule planning such as figuring out how to navigate what courses to take, especially if our path deviates from the formulaic roadmaps our schools may have outlined for us. Whether it’s a delayed graduation because of a gap year or major switching, we all want to have our heads in the right space! We want to feel prepared!

Every semester, we can find it difficult to get in immediate contact with an advisor, and even when we do, a majority of students consider their experiences unhelpful regarding how to move forward with enrolling in classes.

What it does & how we built it!

Frontend: For the frontend of the stack we use the Javascript library "React.js" to create this user-interactive components and to format the interface we utilized CSS. We wanted the first landing page to be the student's chat with the 'AI Advisor' through "Builder". Through here the user inputs texts or pdfs and the AI (backend) parses through everything to stores the students information about courses taken for their major. We also have "Courses" and "Schedule" all on the left hand side so users can also manually input courses as well as schedule where the AI spits out the schedule for the semester.

Backend: For the backend, we used Python and some of the libraries and APIs offered such as Pydantic and Open-AI (and a lot more!) to create the actual functionality of the web app. This side parses through the pdf with Fitz and also parses through the texts in order to store it as a JSON. This then checks for the academic roadmap of the school to align with the roadmap of the student, see which classes they have taken and verify what they can take next with prerequisites or if it is offered. This will then loop through, with how man credit hours they can take, assign the respective amount of classes for each semester or just create a schedule for the current semester.

Challenges we ran into

A majority, if not all, of the languages we used is something we had to learn on the spot which is especially difficult to learn within less than 24 hours. Our team was also originally supposed to be 4 people with roles assigned, but it had ended up being 3!

For the more technical side, we had difficulty working an actual chat AI for the student which is unfortunate especially since that was the purpose for the project.

Accomplishments that we're proud of

We are proud of actually getting something running however, but we are even more proud of the ideation process and understanding the libraries and APIs.

What we learned

We learned a lot about the languages and tool we used to build this, but we also learned a lot about how to manage and delegate tasks to build a web app. Most importantly, the biggest learning curve was learning how to connect frontend to backend or backend to frontend.

What's next for AI Advisor

If we can fix the bugs first, implement databases for all roadmaps, as well as if it is offered for the semester, we can get a very useful and working tool for students. If it is really plausible, we can scale so it can be used for all universities! After fortifying these existing features, we can implement new features like preferences personalized to each student like what classes they may prefer or which professor may be a better fit. We can also standardize a way to submit major declaration and a lot more!