App and Web Development Club officers compete at the University of Pennsylvania Hackathon

Seniors Gautham Raghupathi and Tom Zhang build an AI customer prediction model at the PennApps 20 hackathon.

Gautham Raghupathi and Tom Zhang work on their machine learning project for a hackathon at the University of Pennsylvania. PHOTO BY RAGHUPATHI // USED WITH PERMISSION

Devin Gupta
September 27, 2019

“Artificial intelligence” and “machine learning” are common buzzwords about programming, but the App and Web Development club’s officers of Android development Gautham Raghupathi and Tom Zhang are working to teach MVHS students about common uses of programming.

Recently, they attended the PennApps 20 hackathon hosted by the University of Pennsylvania on a high school sponsorship, where they competed against undergraduate and graduate students to build solutions to common financial problems, such as projects that predict stock prices. 

From left to right; Tom Zhang, Iyad Okal, Gautham Raghupathi, and Amy Yuan. PHOTO BY RAGHUPATHI // USED WITH PERMISSION

For their project, Raghupathi and Zhang built an algorithm in 48 hours that uses cameras in a business to predict when and in what volumes consumers will appear. Then, using analytics, the algorithm determines peak hours and when to run promotions.

“[Our] goal is to help businesses understand customers better, and to help them improve their performance and efficiency.” Zhang said.

As an owner of a business who might use this platform, the General Manager of Cupertino Marketplace, a convenience store on Stevens Creek, Tammy Jones describes ways she might use this project in her business.

“It [the algorithm] would give me the opportunity, with the predictive [predicted value] to plan not only staffing but food needs,” Jones said. “So predictive in terms of [meals required for] breakfast versus lunch versus dinner.”

The Raspberry Pi Zhang and Raghupathi used to build their project on. PHOTO BY RAGHUPATHI // USED WITH PERMISSION

Building a machine learning model that can identify and predict customers is a project many large tech firms like Google and Amazon work on, according to the New York Times, but these students did it in a matter of days using online web tools. They used a Raspberry Pi, which is a computer commonly used for robotics. They also used Google API, a platform developed by Google that allows people to utilize Google’s AI recognition technology.

“We used Google API for [the camera part],” Raghupathi said. ”That was easy to do… Then, after that, we built an SVM [support vector machine] that uses all the variables [necessary]. And as you train it… you can start predicting who is going to come tomorrow.”

Gautham’s sister, sophomore Chaitra Raghupathi, also comments on her brother’s work ethic and motivation to complete this and other projects.

“He’s always been really driven, but I think he really started pushing himself in sophomore year,” Chaitra said. “[Having Gautham as a brother is] really encouraging actually because when you’re surrounded by really smart people, you always want to do better yourself.”

Gautham and Zhang teach what they learned at the App and Web Development Club that meets every Tuesday during lunch in E202.