Although he’s been teaching IEOR courses at Berkeley for half a semester, Professor Daniel (dah-NEE-yell) Pirutinsky has no clue what his office looks like, what campus looks like, or even what Berkeley looks like. Hired in the midst of a global pandemic, Pirutinsky has spent the entirety of his nascent career as an assistant teaching professor in the IEOR department from his home in New Jersey.
As unconventional as the start of his teaching career has been, Professor Pirutinsky’s journey to Berkeley is even more surprising.
Raised in New York among an ultra-orthodox Jewish community, Pirutinsky never expected to attend university, let alone become a professor. While studying at a religious school, Professor Pirutinsky found escape in work, and “bent the rules” by finding various jobs in the area, where “after about three months, [he] ended up improving how each business ran without even realizing it.” As a young queer person, Professor Pirutinsky was challenged by the claustrophobic environment of religious school and the grueling 7AM-9PM days to which he was subjected. As he continued bending the rules, one day in the local public library, he happened on a career book introducing him to operations research. After some reading, Professor Pirutinsky realized that he had already been optimizing and improving the businesses he worked at for months; conducting operations research was already second nature.
Operations research seemed to be a perfect fit for Professor Pirutinsky, except for one critical obstacle. As any IEOR/ORMS student can attest, operations research demands a strong working knowledge of mathematical concepts — a subject Pirutinsky had always despised. Eventually, however, he decided to “give math a fair shot,” and spent two and a half years of lunch breaks teaching himself math, beginning with algebra and working through trigonometry and calculus. After years of self-teaching and diligent practice, Professor Pirutinsky had developed a healthy relationship with math; coming to understand “why it’s important, why it’s challenging, and why it’s ok that it’s challenging.”
Identifying operations research as his field of interest and self-learning the requisite math, were not all that needed to be done. Professor Pirutinsky faced another daunting obstacle: obtaining an undergraduate degree and preparing for graduate work in the field. He was expressly forbidden to attend traditional college and was strongly discouraged from pursing any form of higher learning that was not dedicated to religious texts. Pirutinsky’s solution was to enroll in an accredited test-for-credit program, through which he studied and took tests in his limited free time. As Professor Pirutinsky remembers, “I [earned my degree and prepared for graduate study] without my community’s blessing, but they couldn’t really object since I was still in religious school at the time. It’s amazing that I was able to pull it off.”
From there, Professor Pirutinsky took a giant step forward and, “with the help of many kind people,” was able to enroll as a PhD candidate in Operations Research at Rutgers University, a degree he earned just this year.
Professor Pirutinsky’s research interests span a wide range of topics, but mainly center around reinforcement learning. Modeled on the way humans learn, reinforcement learning employs positive and negative stimuli to encourage computers to express certain behaviors. Just like the pain of touching a hot stove uses a negative stimulus to encourage you to avoid such behavior in the future, Professor Pirutinsky hopes to apply the same principle to train artificial intelligence to behave in specific ways.
An example of a reinforcement problem that Professor Pirutinsky has explored during his academic career is the bandit problem. The simplest version of the problem describes a number of slot machines (whose nicknames ‘one-armed bandits’ gives the problem its title) and one gambler, who seeks to learn which machines are the best investments while minimizing losses from playing. According to Professor Pirutinsky, “the bandit problem is, at its simplest level, about exploration vs. exploitation; how to learn something worthwhile about the world while minimizing risk and maximizing reward.”
While Professor Pirutinsky is mainly interested in understanding the gap between theory and application of reinforcement learning problems, potential use cases abound, from A/B testing of online advertisements to determining which COVID-19 therapies are most effective while achieving optimal patient outcomes.
Although Professor Pirutinsky’s tenure at Berkeley has only just begun, he offers three pieces of sage advice to any first year students who, like him, may find remote work in a new community strange and challenging. First, says Pirutinsky, be kind to yourself. “This isn’t a regular semester, nor a regular year. You may feel stressed and worried and have barriers in the way of your learning, so accept those and take care of yourself!” Second, prioritize social opportunities and keep your camera on. “We’re social creatures, and it’s important to feel connected to our peers and colleagues.” Finally, Professor Pirutinsky reminds everyone to remember that anyone they might interact with, “be they students, staff, or faculty, are also living through 2020. Be kind to each other, and remember that this isn’t a normal time for any of us.”
While teaching IEOR 172, 242, and enjoying healthy doses of Netflix have kept Professor Pirutinsky busy through his first semester at Berkeley, he looks forward to someday “finding out what [his] office looks like.” Despite missing the benefits of an in-person semester, Professor Pirutinsky is optimistic. “It was only a few years ago that I was feeling trapped in my religious environment, with no hope of attending college or pursuing my interests. In perspective, it feels like the world has only recently opened up to me.”
Daniel Pirutinsky is an Assistant Teaching Professor in the Industrial Engineering and Operations Research Department at UC Berkeley. Please reach out to Matt LoPolito (firstname.lastname@example.org) with any questions or comments.