**team members and AI disclosure will be in project repo readme file I was inspired by a YouTuber named Gohar, who makes videos about surviving your first year of school. In one of his tips, he mentioned how knowing your school's syllabus can help you organize your time and prepare for the curriculum. However, now that I'm in university, I realized that syllabus are often very long and filled with a lot of information that is often irrelevant, which made me lose track of the truly useful details.
That's how Syllabus Unifier was born, a web app that takes the syllabu from all your subjects as input and returns a PDF summarizing everything important, such as dates, evaluation criteria, professor's contact information, etc. In addition to that, if you add a file with your schedule, it gives you an .ics file so you can import your classes into Google Calendar and have them organized.
I used Gemini AI to give me a prompt that I could pass to Copilot to help me develop the backend prototype. This is built in Python using libraries that read the documents, extract the important information, and return it via an API call from the frontend. The frontend was developed by me using React js in combination with React Blitz and AI to correct minor errors and style the page.
There were quite a few challenges, especially using AI, as it would lose track when the files became very large or the prompts too abstract. It would make mistakes during debugging and even force what it believed was the solution, even when I only asked for recommendations.
Even with all this, the experience was very rewarding, and I leave proud and happy because I had to understand the code the AI generated to correct FUNDAMENTAL errors it had left behind by writing code with errors like not reading or getting stuck on non-existent variables.
Thanks to this experience, I've gained a lot of knowledge, especially in using AI to make my work more efficient, analyzing its code, and debugging instead of just letting it do everything.
Finally, I would like to refine the algorithm for extracting dates, as it sometimes has problems, and I would likev this project to become a reliable tool for students.
Log in or sign up for Devpost to join the conversation.