36 hours. Teams of four from across Maryland. One track.
A coding project that aims to challenge undergraduates to come up with creative solutions and present them to a panel of specialized judges.
HopHacks — a competition that brings students passionate about computer science and engineering together — embodies all this. The competition fosters broad-scale innovation through its many tracks, from the open-ended general track to more specialized tracks such as the Patient Safety Technology Challenge and the Philanthropy Track. A member from one of this year’s winning teams provided a new perspective on the HopHacks experience in an interview with The News-Letter.
Sophomores Tom Wang, Andy Shi, Lily Ru and Harry Jiang competed in HopHacks this year on the philanthropy track. Their team developed an artificial intelligence (AI) powered note-taking tool geared to help students with disabilities.
Wang discussed the idea-generation process and explained how the team came up with their idea.
“Harry spotted an email from SDS that said that they were hiring student note-takers for… a lot of classes for students with disabilities. That was also a spark point that made us want to solve this issue with AI,” he said. “With this idea, we all felt that if we were to make [this tool], we ourselves would be willing to use it.”
After the first few hours of brainstorming, the team delved into planning their code and the features they wanted to include in their software. One of the biggest things they focused on was differentiating between must-have features — essential parts of their design — and nice-to-haves that could improve on the basic skeleton of their design.
Wang credited his own research experience under Dr. Jason Eisner as the source for many of the Natural Language Processing (NLP) techniques that helped him and his team in coding their project. These methods especially helped in ensuring that the team’s software was multimodal — capable of integrating scribbles, audio files and pictures of blackboards to create comprehensive AI-generated notes. However, integrating this multimodal approach was one of the key challenges the team faced during the development process.
“We spent a lot of time trying to get the audio-to-text to work, and after a while, we reached a limit for the service that we’re calling…suddenly the fundamentals were breaking down,” Wang explained. “That’s where we needed to pivot to another solution, and [that’s why] I’m glad we separated things. Even though this specific block was failing, it wasn’t affecting other things, so we could just separate it and fix it. In the end, we turned to another service like OpenAI and used their API calls, and that solved the issue.”
The team's other major challenge was in the front end, or how information is displayed. To ensure that the display of their project was just as sound as the backend work or the actual technical features, the team split into two parts initially.
Two members worked on the frontend and two on the backend until it was functional, at which point all four team members came together to focus on the frontend, ensuring that what they built was visually appealing in addition to being fully functional.
“We’re building an app for students, not just a service for companies. User experience will change a lot. If you have a fully functional backend, but your frontend is not that beautiful to work with, it won’t have this good effect,“ Wang said. “In terms of importance, frontend is definitely very important, especially in the general hackathon atmosphere. For hackathons, you want to display something that attracts people, so that’s why we spent a lot of energy on it.”
Wang also provided advice for prospective hackathon competitors.
“Build something you want to build. Don’t build it for the hackathon; don’t build it because it’s fancy …just ask yourself, ‘If I build this out, will I use it and enjoy it and make it part of my life?’ That’s the question we asked ourselves during this year’s hackathon, and our answer is yes… we found this tool would be essential for us, so that’s why we wanted to do it,” Wang concluded.
