From promoting Pi Day on live TV to penning articles in Scientific American, Max Springer (Ph.D. ’25, applied mathematics & statistics, and scientific computation) brings math to the masses. 

If you’ve ever wondered about artificial intelligence (AI) chatbot hallucinations, the accuracy of tornado science in the 1996 film “Twister” or whether the “stable marriage problem” in mathematics could yield better matches on the reality dating show “Love Island,” just ask University of Maryland recent graduate Max Springer (Ph.D. ’25, applied mathematics & statistics, and scientific computation).

Through a 2024 Mass Media Science & Engineering Fellowship with the American Association for the Advancement of Science, Springer wrote about mathematical topics and more for Scientific American, a publication he devoured for years but never dreamed of writing for. “I grew up reading Scientific American,” he said. “My grandpa got me a subscription when I was in middle school and I never really stopped.”

Springer doesn’t plan to pursue a career in journalism—instead, he will study fairness in algorithms as a postdoc at Princeton starting this September. However, a grounding in science communication helped Springer break down complex subjects and share his passion for math with a wider audience. Most recently, Springer joined Mathematics Professor Larry Washington for a Pi Day interview with Baltimore-based station WJZ-TV. The two mathematicians explained pi’s relevance beyond its connection to sweet treats, and in doing so, showed that math can be accessible—and even entertaining.  “People think that math is so abstract and hard to understand,” Springer said, “but when you write a good story about it in layman’s terms, readers quickly get drawn in.”

‘A beautiful merging’

Springer’s appreciation for popular science started at a young age. He recalls looking up to astronomer Carl Sagan and Bill Nye, the TV host who introduced a generation of children to the fun side of science, technology, engineering and mathematics. “Those were the first people who taught me about science and, above all else, to be curious about everything,” Springer said.

At times, Springer’s fascination with “everything” made it challenging for him to find his niche. During his sophomore year at Cornell University, a professor helped Springer pick a major as the declaration deadline loomed. “He didn't tell me to pursue math, but he emphasized that you can explain so many different phenomena using mathematical principles,” Springer said. “Studying math was a beautiful merging of all my interests.”

That advice ultimately sealed Springer’s decision to pursue a career in applied math. After graduating with his bachelor’s degree in mathematics in 2019, Springer completed a yearlong research position at Yale’s medical school, where he used machine learning and statistical methods to predict when patients with epilepsy might experience seizures. While looking for a graduate school that would continue to expand his diverse interests, Springer felt that UMD’s interdisciplinary applied mathematics & statistics, and scientific computation program would offer the best variety. 

The price of fairness 

After joining UMD in 2020, Springer started conducting research with Professor Mohammad Taghi Hajiaghayi, who holds the Jack and Rita G. Minker Professorship in UMD’s Department of Computer Science. Hajiaghayi introduced Springer to combinatorial problems in algorithm design, including how to efficiently and equitably divide resources among people. “That got me interested in the idea of what it means for an algorithm to be fair, and conversely, what does it mean for an algorithm to be biased,” Springer said. 

Springer started analyzing clustering problems, or ways of grouping similar data points and discovered that these technical considerations could have real-world impacts. This included cases where the data used to create clustering algorithms contained underlying biases. “There are a lot of instances where clustering algorithms are used to decide mortgage or which ads to deliver to people,” Springer explained. “What's been shown countless times in the past decade is that these algorithms can inherently pick up on people's gender or race and discriminate on those grounds.”

Springer’s dissertation focused on one central question: Does incorporating certain fairness constraints into an algorithm make its solution—or outputs—less accurate or less optimal? Ultimately, Springer found that “the price of fairness” was significantly lower than he expected. “What I showed across a bunch of problems is that when you incorporate these fairness considerations, you don't degrade your solution by much,” Springer said. “I hope that my research serves as a statement and shows that we should be considering these problems through this lens.”

While engaged in these technical problems, Springer also thought about how to communicate his work—and science more generally—to non-experts. When he first started writing for Scientific American last summer, he struggled to switch from an academic style of writing to popular science, which aims to break down complex subjects as clearly as possible.

“I wrote an article about satellite debris collecting in the ozone layer, and the physics editor marked up my entire draft and said, ‘You’re writing like an academic,’” Springer said. “I finally realized that if you’re talking about a research result, you want to walk the reader through that discovery and show them that science is a natural pursuit and they can understand it, too.”

Responsible AI

Another one of Springer’s interests is a subject that attracts widespread attention: the ethics of AI. “The AI space is moving so quickly, especially with large language models,” he said. “A lot of my more recent research was about how we can steer these models to ensure that AI systems don't develop to be biased in any way.” During his last year at UMD, Springer worked on data efficiency problems related to machine learning as an intern with Google Research’s Algorithms and Optimization Team. He also served on the inaugural student committee for the Association for the Advancement of Artificial Intelligence (AAAI), an international scientific society aimed at advancing safe and responsible AI that benefits society. 

“Graduate students are the backbone of AI research, especially in academia, so there are a number of things that we wanted to work on,” Springer said of his work with AAAI. “In particular, I wanted to work on mentorship and developing a host of resources for students who are interested in AI research.”

Springer helped to coordinate the AAAI’s first hackathon, in which competing teams developed “cutting-edge AI solutions” over a weeklong virtual event in February. Harking back to his interest in science communication, Springer also supported the development of an AAAI podcast series with student hosts interviewing AI experts. As Springer reflects on his time at UMD and starts packing his bags for Princeton, he feels that familiar sense of curiosity propelling him forward.

“When I read an article or a preprint that comes up, I’m instantly generating research questions,” Springer said. “I think that’s probably a good sign that I’m ready to start defining my own research. It’s super exciting.”

Written by Emily Nunez