Going for Goal

Shio Huang (MLDS '25), Wuyou (Tony) Shu (MLDS '25), and Zeyu Zhang (MLDS '25) were first-quarter students when they were introduced to Northwestern Engineering's Master of Science in Machine Learning and Data Science (MLDS) program's annual hackathon. 

In 2024, the trio collaborated to help Northwestern's men's soccer team understand how it could leverage AI in its day-to-day training and operations. Shio, Tony, and Zehu finished third in the competition.   

When it was time to sign up for the 2025 hackathon, the group had a singular goal: win first place.  

That focus paid off.  

Shio, Tony, and Zeyu won the 2025 hackathon, which was sponsored by the Chicago Fire of Major League Soccer. The team challenged students to take a public soccer dataset that contained transfer data, player valuations, and match event data and in 48 hours develop solutions predicting player valuations, identifying breakout candidates, and analyzing player development trajectories. The competition was open to students across Northwestern. The winning team will tour the team’s performance center and receive tickets to a home game next season at Chicago's Soldier Field. 

Shio, Tony, and Zeyu created an intelligent, fully automated scouting platform that combined machine learning, time-series analytics, and multi-agent generative AI to produce professional-grade player reports instantly.  

"Our solution stood out because we combined solid traditional modeling on tabular data with a multi-agent LLM layer that pulls in live web information for social impact," Zeyu said. "On top of that, we were able to automatically generate a clear, tailored report and share it through an interactive chatbot, so stakeholders not only get predictions but also an easy way to understand and explore the insights." 

That understanding aspect stood out to David Portugal, software engineer with the Fire. 

"Their use of a conversational AI layer that sat on top of their AI-generated player reports impressed me because their work not only provided meaningful insights but also had real business value," he said. "Giving the user the ability to search for and compare players conversationally has tremendous business value as it closes the gap between the technical, analytical side and the non-technical business side. This is incredibly important when working with non-technical stakeholders." 

Tony explained that he and his teammates often found themselves applying many skills they learned in MLDS during the competition, including how to leverage data engineering, generative AI, and predictive modeling.  

"Participating in the hackathon offered a crucial opportunity to implement our skillsets into something meaningful and practical beyond coursework," he said. 

The most important skill the team incorporated was not a technical capability—it was a mental shift. 

"What matters most is not the skills or techniques themselves, but the way of approaching a problem we learned throughout MLDS, which helped us to build a comprehensive knowledge base and amass numerous methodologies," Shio said. "What the Hackathon provides is not only a chance to showcase your take-aways from the lessons, but to foster resilience under a real working situation when the deadline is close. It’s a valuable opportunity to go beyond your role as a student." 

Portugal agreed. Just as importantly, the Fire wanted to provide an opportunity for students to apply lessons learned in the classroom to a real organizational challenge. 

"Working on real business challenges is a great way for students to learn that building a great model is only half the battle," Portugal said. "The other half is making sure it actually gets used—which means understanding the business problem, designing for your users, and effectively communicating value to stakeholders who aren’t always going to be as technically proficient."