Resolve Magazine Fall 2023 >> Making Sense of Machine Learning >> Stories >> Preparing future professionals
Corporate-sponsored projects in machine learning are giving students in Lehigh’s MS in Financial Engineering (MFE) program, like Joao Ji Won Lee ’24G, an edge in the job market.
Working with industry partners, Ji Won Lee says, has built up his experience with technologies like AWS (Amazon Web Services), Azure (a cloud computing platform), and Tableau (a visual analytics platform). “I just had an interview with Bloomberg,” he says, “and that practical experience allowed me to really engage with the interviewer and differentiate myself from other applicants.”
While the program (a joint offering of Lehigh’s College of Business, College of Arts and Sciences, and the Rossin College), doesn’t teach machine learning, students are expected to learn it, says MFE program manager and Lehigh Business teaching assistant professor Patrick Zoro (top photo, right). “If you don’t know Python, you can’t really work on these projects.”
The program has more than 30 industry-supported projects, some of which are part of Lehigh’s Mountaintop Summer Experience. Ji Won Lee is involved in two such efforts. In the first, he employs reinforcement learning to build a user-friendly dashboard for both professional and casual traders to help them make better-informed decisions around buying or selling certain assets. In the second, he’s using a natural language processing model—one that combines computational linguistics, machine learning, and deep learning models to process human language—to categorize companies in a way that can help portfolio managers identify those that are giving out high returns.
In another Mountaintop Project, MFE student Sisheng Liang ’25G works on the interpretability of supervised learning. Essentially, making models designed for traders more explainable.
“Right now, there’s a trust issue with some of these models,” he says. “Traders don’t understand how they work, so they aren’t always willing to use them.”
Zoro says such hesitancy makes sense—especially when millions of dollars are at stake. “Interpretability is actually a word now in machine learning,” he says, “and the ability to interpret findings is itself a kind of science. Otherwise, these models are just a black box.”
Liang came into the program with zero experience in machine learning and interpretability. But in addressing real-world challenges, he’s become adept at searching out and acquiring the skills he needs to solve the problem, and that will set him up for a career as a quant (a financial industry professional skilled in advanced mathematics and computing).
“I’ve learned Python and how to code,” he says. “And because so much of what we need to know isn’t in a book, I’ve learned to source the information I need from Google, GitHub, YouTube—so many different places. It’s really satisfying to know you can solve these specific problems in industry.”
Main image: Christa Neu
