John Silberholz: Magnet to Markets

by Gracelyn Chen '28 for Silver Quest

Silberholz traces this mindset to his SRP (senior research project) at Blair, where he was first introduced to the field of operations research. Though he describes it as “much more obscure of a field than math or computer science,” operations research stood out to him because it combined his existing interests in math and data. 

During his SRP, Silberholz worked on optimization problems that focused on finding the best possible decisions under constraints, an approach that has shaped much of his work ever since. “There’s a data aspect,” Silberholz explains. “There are constraints, there are limits to what we can or can't do. We want to do as well as possible based on our understanding.”

Drawn by the field’s many algorithmic foundations, Silberholz continued pursuing operations research beyond Blair, eventually centering his PhD and research around data-driven decision making. “That’s what I ended up doing my PhD on,” he says. “All of my research that’s followed has been quite closely connected to this idea that we want to build a model of some of reality as best we can, and then optimize decisions.”

In terms of his teaching, Silberholz asserts that these same principles apply. As a business professor who frequently advises PhD students with prior industry experience, he finds that much of what his students learn ends up being very applicable to real-world business situations. “It’s really rewarding,” Silberholz says, “Many [students] are already coming in thinking of business experiences that they’ve had and seeing how some of the tools that they’re learning in the coursework directly apply to those sorts of situations.” 

Silberholz frequently teaches what he refers to as “big data” courses. “To me, big data simply means data where somehow your normal way of performing data analysis to build a model, and maybe gain insights or make decisions, is going to break down,” he explains, “Either because the data are indeed so big that you don’t know how to fit the models you’re used to … or your data are text or videos or some [unfamiliar] format.” In these courses, students learn to extract value and meaning from such datasets. 

Silberholz emphasizes that the true value of these classes lies outside the classroom. “Definitely one of the most rewarding things,” he says with a laugh, “is when a student reaches out and tells me about some random application where they ended up using the skills.” 

When he’s not teaching, Silberholz turns his attention to research, working on various projects that pique his interests. One such project involves a pilot study conducted with collaborators at the University of Pennsylvania, in which the team explores the best ways to encourage participants to improve their health by increasing their daily step count. Rather than relying on the same strategies to motivate everyone, the study uses a method known as dynamic adaptation. “We try to learn dynamically, as we’re interacting with participants, the sorts of outreach that are particularly effective for them,” Silberholz explains. “[That way] we can allocate limited resources … where it’s going to be most likely to get them on track.” He notes that this type of information is especially valuable for health providers and insurance companies, who fund programs promoting healthier lifestyles and need to spend their resources where they will be most cost-effective.

Silberholz recalls enjoying his time in the Magnet. “It’s hard to think of a class I didn’t enjoy,” he says. Though he knew early on that math and computer science were his main interests, Silberholz made sure to explore the program’s many other offerings. “I felt like Blair had so many cool-sounding opportunities in biology, in physics, in engineering, that I found myself kind of unable to avoid them.” He doesn’t mind the lack of academic focus in high school. 

Silberholz encourages current Magnet students to take advantage of Blair’s wide variety of courses and explore broadly while they still have the opportunity to.