My team recently won the SP Jain Gen AI Hackathon. We were overjoyed to win the grand prize, which was an Amazon gift card worth 500 Au$, but that was just the cherry on top, because the journey of building our solution alongside my amazing team members, Abdualim, Naufal and Maleakhi, was immensely rewarding.

The Challenge 
We were tasked to build a web-based chatbot named ‘Schedule Pandit’, which helps students stay on track with their schedules. The chatbot is supposed to handle conversational queries, respond to student requests, and maintain a respectful dialogue, making it accessible to students of all backgrounds. University life can get hectic, and I know first-hand how helpful it would be to make checking schedules easier. 

Our Approach
From the beginning, our strategy revolved around breaking down the problem into manageable components: scraping data from SP Jain Global’s Sydney schedules, integrating the GPT-4o mini API and designing a web-based UI for Schedule Pandit. With a strong foundation of coding, machine learning, and a little bit of UI design that we learned in and outside of our classes, my team’s background as BDS students proved invaluable as we tackled these challenges one by one.

We knew that a lot of our time was going to be used on scraping the data, since the schedules are structured in a way that cannot be accurately read by ChatGPT. This meant that we had to find a way to format these schedules in a way that is easily readable by the program. There was a lot of discourse regarding our approach, but we settled with “translating” these schedules manually into a simple text file which Schedule Pandit could read easily. Abdualim and Maleakhi came up with the Python code to interpret these schedules, which streamlined the whole process by a lot. I tinkered with the implementation of the GPT model, so that we ended up with a Schedule Pandit which gives respectful responses and asks clarifying questions to give the most accurate responses possible. Naufal was the star of the show when it came to the GUI, making sure that it was as clean and user-friendly as possible.

Working as a Team
Collaborating with Abdualim, Naufal, and Maleakhi was a highlight of the hackathon. Having worked together on our Capstone Project, we already had a strong communication dynamic, which made the whole process smoother. I believe each team member brought valuable insights that strengthened the chatbot, and our combined efforts were the key to success.

What we learned 
Taking part in the Gen AI Hackathon taught us some valuable lessons. By blending teamwork and technical skills, we were able to turn our idea into a practical, user-friendly tool. Winning the competition has motivated us to keep exploring new possibilities with AI, and Schedule Pandit is just the start of what we hope to accomplish!

-Oliver Y.
BDS Batch 2022