As an undergraduate physics major, Travis Brooks participated in advanced research at Fermilab. “It was eye-opening,” he says, “to see what happened when you had to interpret data coming out of an actual collider.” Photo by Jason Doiy.
In the digital era, we’re used to being able to find just about anything online, usually starting with a search. Type in a few keywords, and boom—results fill the screen. Simple, right?
That instant gratification belies just how complex and powerful search algorithms have become since the early days of the web. Today, online searches are driven by a growing array of variables, including where you are, what device you’re using and what you clicked on two hours ago.
That’s why tasks as seemingly simple as picking which show to watch tonight are driving a growing field full of software engineers, machine learning experts, data scientists—and at least one former theoretical particle physicist, Travis Brooks (CWR ’97). As a data science and engineering leader at Netflix, Brooks heads a team that focuses on improving members’ experiences on the entertainment service.
In one sense, the job is a world away from the one Brooks imagined when he graduated from Case Western Reserve, bound for Stanford University and an advanced degree in theoretical particle physics. But there’s a common thread, and it unspools from a remarkable project during his days at CWRU.
In high school, Brooks was a polymath. One of his English teachers thought he should become a historian or literary scholar. But instead he gravitated toward the logic of math, and he wanted to explore the deep mysteries of the universe. So once he started college, he decided to major in physics.
“I loved working at that interface between the purity and freedom of mathematics and the constraints of the real world,” he says. “I thought that’s what I wanted to do with my life.”
During his sophomore year, Brooks was invited to join an experiment called MiniMax at Fermilab’s Tevatron collider in Illinois. Cyrus C. Taylor, the Albert A. Michelson Professor in Physics and one of Brooks’ teachers, organized the experiment with Stanford theoretical physicist J. D. Bjorken to search for laser-like emission of subatomic particles in grazing high-energy collisions. The goal was to detect particle interactions that were (and still are) purely theoretical.
The MiniMax project was an outlier on several levels. One of only three collider experiments at Fermilab in the 1990s, it was designed to take advantage of a quark data-acquisition capability that was about to be dismantled. With the clock ticking, MiniMax compressed a process that would normally have taken a couple of decades into about a year. As for the budget, a paper by Bjorken later put it this way: “The location of the experiment in fiscal space was near the imaginary axis.” Funding, in other words, was almost nonexistent.
Then there was the team. In particle physics, it was very rare for theoretical practitioners like Taylor and Bjorken to cross over to the experimental realm. Also unheard of for such an experiment: full participation from undergraduates. With a crazy schedule and no money, getting doctoral students involved in MiniMax just wasn’t going to work. But, Taylor says, “I knew how good our undergraduates were.”
His recruitment of Brooks and two other undergraduates, Matthew Knepley (CWR ’94) and Erik Kangas (CWR ’95), was initially met with dismay from some of his colleagues. “But after the students presented to the collaboration,” Taylor recalls, “they were all voted full members. That’s the part that was truly extraordinary.”
MiniMax ultimately did not detect the elusive quantum fluctuations it sought. But Brooks says the project was the highlight of his physics education.
“It gave me an opportunity to see what the practice of physics was like in real conditions rather than in classroom conditions,” he explains. “It was eye-opening to see what happened when you had to interpret data coming out of an actual collider.”
The website for MiniMax, created by participating scientist Jon Streets, was among the first 100 in the world when it went up in 1993, giving Brooks an early, if unintentional, presence on the web. Brooks also emerged from the project with first authorship on some high-profile papers—a feat that very much impressed his scientist mother, he jokes, until she found out that author listings in particle physics journals are alphabetical.
In conversation, Brooks tends to qualify his own comments in true scientific spirit, noting when they are “anecdotal, not a statement from data,” or somehow in need of validation. Growing up in Memphis, Tennessee, he played logic games with his dad, a trained philosopher. His mother was a geneticist and practicing physician. Did he have an unusually intellectual upbringing?
“My sample size is not large on this,” he answers, before surmising that outside observers would probably say yes.
Travis Brooks’ job as a data science and engineering leader at Netflix is a world away from the one he imagined as a physics major at Case Western Reserve. Today, he lives with his family in Berkeley, California. Photo by Jason Doiy.
His father stayed at home to raise him, and Brooks remembers seeing the movie Mr. Mom and “being very confused at why this guy was so incompetent at managing a house. This is what my dad did, basically.” His many philosophy-oriented conversations with his father primed Brooks to follow a discipline rooted in pure reasoning, while his exposure to his mother’s career left him with a desire to have an impact on other people.
After CWRU, as a doctoral student at Stanford, Brooks found himself becoming more interested in what he calls the “service aspects” of his graduate career—joining committees, working on issues affecting the scientific community as a whole—than in his own research. Skipping his dissertation, he took a job at SLAC National Accelerator Laboratory (formerly Stanford Linear Accelerator Center), where he managed a search service called SPIRES for particle physics literature.
“That very much scratched my itch of being in community service while still indulging in my passion for physics,” Brooks says. He stayed for nine years, leading an international collaboration to replace SPIRES with a next-generation knowledge management system for high energy physics. That period coincided with the internet-fueled explosion of tech products around Silicon Valley.
“I started understanding that what I was doing was actually called product management,” Brooks says, “something that a lot of people nearby to Stanford were doing, except at a much bigger scale.”
So in 2011, he joined the business search and review company Yelp as a product manager for search and data mining. He went from working on a product with a user base of about 10,000 physicists to a platform where millions of people a month were searching for everything from the best breakfast spots in town to the top local dentists.
At Yelp, Brooks became familiar with some of the thornier data problems now common across digital enterprises. No longer just an information hub, the web has evolved into an experience engine where people shop, socialize and stream entertainment. That makes the basic notion of “search” a lot more involved than people might think.
Let’s say you type “oscar” into your web browser. Chances are you’ll see recaps of this year’s Academy Awards in the top results. But what if you’re looking for the health insurance company, or the judicial hiring portal, or the 1991 movie starring Sylvester Stallone? Here is where data science comes in, making inferences about what you want and predictions about what you will want.
A few years in at Yelp, Brooks realized he needed data specialists. He had plenty of engineers who were competent in using statistics to figure out how people were using the service, but no full-time experts. So he hired one, and then another. Eventually, he was running a team of about 25 data scientists and several product managers, all of them working to make Yelp’s search function, recommendations and overall product even more tailored to people’s needs. Then, last summer, he took the job with Netflix, which offered him a chance to build on his work using data to understand consumers.
Physicist colleagues had warned Brooks when he entered the tech sector that it would be boring. Filled with corporate drones. Not as free. “To me, it’s been exactly the opposite,” he says. “People who have to use the tools of mathematics within the constraints of reality—an ever-changing reality of different users and different technologies every quarter—actually get very innovative, very quickly. It’s an incredibly dynamic and fun environment because of that.”
In the years since his graduation from Case Western Reserve, Brooks has remained connected to the university. While at Yelp, he returned to campus to meet with students at career fairs—and sometimes he hired them. In 2016, he accepted an invitation from Taylor, who was then dean of the College of Arts and Sciences, to join the college’s Visiting Committee, an advisory board made up of distinguished alumni. “Travis is one of the younger members of the committee,” Taylor says, “and his vantage point on the tech side is especially helpful.”
Timothy Beal, the Florence Harkness Professor of Religion and chair of the Department of Religious Studies, has also benefited from Brooks’ insights. Beal, who served as interim dean during the first eight months of 2019, has long explored the application of data science to humanities scholarship. Along with Institute Professor Mark Turner in the Department of Cognitive Science, he is leading an initiative to advance data science in the college. Brooks’ views, Beal says, have significantly influenced his own thinking about the future of data science, both in the tech sector and in universities.
“As Travis put it to me at a Visiting Committee meeting, too much of data science is really ‘data engineering’ rather than scientific pursuit,” Beal recalls. “We need to leverage big data and machine learning to ask new research questions that can advance knowledge and understanding.”
Brooks says that several companies, including his current one, do a good job of thinking in a hypothesis-driven way. “But I still think we have room to grow as a community of data scientists,” he adds. “Are we really scientists—are we really finding more general principles? That’s a tough question that I haven’t seen people answer as much as I’d like.”
Brooks’ perspective is also valuable to the college, Taylor says, because his professional trajectory illustrates a general truth about the experience of today’s alumni: “They often think, at the early stages of their careers, that they know where they’re headed and what they want to do. But the world changes so fast. Travis has found opportunities he never dreamed of. In fact, he’s doing a type of work that no one dreamed of when he was an undergraduate.”
Christina Nunez is a freelance writer based near Washington, D.C.
Travis Brooks’ mother, Jewell C. Ward, is a professor of pediatrics–genetics at the University of Tennessee Health Science Center. Recognizing the obstacles she overcame as a female scientist and physician, he has honored her with a gift creating an undergraduate scholarship at Case Western Reserve. Photo by Brandon Dill/Memphis Business Journal.
Travis Brooks recently established the Brooks–Ward Family Scholarship Fund at Case Western Reserve in honor of his mother, Dr. Jewell C. Ward. Ward, a geneticist and clinical physician, inspired her son to pursue science and find ways to use it in service to others. Still, Brooks says, “her care and devotion to her patients and her laboratory remain an impossibly high standard.”
The scholarship will be awarded to admitted or enrolled students, with a preference for women in quantitative sciences such as physics, math, statistics and computational biology.
Brooks, who lives with his wife and 7-year-old daughter in Berkeley, California, notes that both of his parents pioneered gender equality. While his mother followed her profession, his father stayed home to care for him. “I’m fortunate to have had generations before me normalize the idea that a woman could have a PhD and an MD and be the scientist and the doctor in charge, and a man could raise a child well,” he says.
Ward began practicing medicine and science at a time when few women were doing so. She routinely wore a stethoscope even though she didn’t need one for her work, because otherwise she might have been mistaken for a nurse. Brooks hopes the scholarship will help future generations of women become doctors, data scientists and researchers who will be recognized on their own terms.
