# Harvard Statistics Interviews Subhabrata Sen about Associate Professor Promotion August 31, 2026 ![Subhabrata Sen](/sites/g/files/omnuum10116/files/styles/hwp_1_1__360x360_scale/public/2026-06/SenS-resized-directory-9-8-2023.jpg?itok=TMeLOfQk) On June 3, 2026, the Harvard Department of Statistics [announced the promotion of Subhabrata Sen](https://statistics.fas.harvard.edu/news/2026/06/subhabrata-sen-promoted-associate-professor) to Associate Professor. In recognition of his promotion and accomplishments, Prof. Sen received the Clark Award from Dean Hopi Hoekstra. As a leading theoretical statistician, Prof. Sen is recognized for his research program that connects probability, statistics, and machine learning, bringing conceptual unity to areas that are often studied separately. To learn more about what motivates Prof. Sen’s research, teaching, and mentoring, we reached out to interview him (please see the edited and excerpted interview below). ## Career Path: 1. **At what point in your education did your decision solidify that statistics as a discipline and academia were for you?** **Sen:** In high school, I liked math and was in luck because the Indian Statistical Institute (ISI) was situated in the city where I grew up, which made me familiar with statistics as a discipline to study. In hindsight, I realize that I didn’t know that much about statistics before joining ISI, but I really loved what I was learning there and decided to stay for the long haul. Towards the end of my undergraduate program, I was exposed to research in statistics and probability and decided that I really wanted to continue on to a master’s and PhD. During my PhD program, I saw the day-to-day life of academics from interacting closely with my mentors, and this resonated with me. I determined that academia was the best place to pursue my curiosity. ## Research and Collaboration Highlights: 2. **Your research spans statistics, probability, and machine learning.** **Describe how your work has evolved over time, including some of your recent highlights.** **Sen:** A unifying thread for my research is a desire to develop new statistical algorithms and to understand precisely what common algorithms are actually doing. During my PhD, a lot of the problems that I worked on were motivated by this sort of practical question. For example, imagine you observe some online social network, like Facebook, and would like to identify similar groups of users. One strategy would be to take the Facebook network and divide it up into chunks of users who are closely interacting with each other. This algorithmic problem raised a number of questions, including: what is a reasonable algorithm to use? What is the best strategy to follow? More recently, I have been interested in three research directions that involve looking at the behavior of algorithms that are used to train statistical and AI models on large datasets. These research directions include examining variational inference algorithms, researching the dependence of algorithms on the underlying data distribution, and evaluating treatment effects in causal inference when there are interactions. In causal inference, for example, I’ve spent the last couple of years thinking about how to evaluate and identify useful algorithms when there are underlying interactions between the units being studied. This problem comes up frequently in the social sciences when policies are rolled out. Because social interaction can influence people, even individuals who aren’t receiving a treatment can have their behavior affected. I’m interested in looking at how and when to use certain algorithms to tease out the causal effects. Moving forward in my research, I am generally really excited to see how my colleagues, students, and I can leverage modern AI tools to solve challenging problems with practical applications. 3. **Your work is also very collaborative. Describe a recent collaboration that you found to be rewarding and impactful.** **Sen:** I have had an ongoing collaboration with Yue Lu, a close mentor of mine who is a professor in Applied Math as well as an affiliate faculty in our department. I was excited to work on a project on spectral methods for multimodal data that we concluded this summer with Statistics PhD student Xiaodong Yang. Multimodal data refers to settings where the scientist observes data from multiple sources on the same entity. For example, imagine you are analyzing people’s interactions over different social media platforms. Traditional algorithms used in a setting with a single data source often don’t work well when combining data from multiple sources for signal estimation. In response, we developed a new spectral algorithm that, under certain well-defined statistical models, is the right strategy to employ. The other project that has been very exciting for me has been co-creating the Stat, AI, and Math seminar with Yue in Applied Math, Horng-Tzer Yau in the Math Department, and Mark Sellke and Pragya Sur in the Statistics Department. The seminar has helped us build a sense of community across departments and schools and has been a venue to keep up-to-date on research breakthroughs in our areas of common interest. ## Teaching and Mentoring: 4. **Having earned the Roslyn Abramson and Extraordinary Teaching awards, what specific classroom practices have created a positive learning experience for your students?** **Sen:** Students at Harvard are extremely motivated and ambitious, so it is always a pleasure to teach them. Most importantly, interacting with students has allowed me to understand their perspective and appreciate that they balance many challenges—a heavy course load, additional research, and extracurriculars. I have a deep respect for students and what they accomplish during their stay at Harvard. In terms of how I teach, I try to present the different concepts as part of a greater whole. When I teach distinct topics in class, I try to emphasize how these topics connect, and more broadly, how the material relates to topics from other statistics or STEM courses at Harvard. Often these ideas can feel disjointed if you just read a textbook. Making these connections in class is important because it motivates students’ desire to dive into the details to better learn a topic. 5. **Several of your students have received the Hoopes Prize and Dempster Award. Through your mentorship, you have exposed many students to research.** **How would you describe your approach to mentoring students?** **Sen:** One of my favorite parts of my profession is collaborating with undergraduate and graduate students. Through my mentoring experiences, I have really come to appreciate the social aspect of research. Thinking together about the same questions, trying out distinct strategies, and sharing the excitement and frustrations are what make research interesting. When I work with undergraduate students, it is often one of their first research experiences. So, I try to focus on the basics of “doing” research, including focusing on core skills such as how to think of a new question and how to communicate their ideas. With graduate students, I can generally be more ambitious. The goal is to identify a general research area that is relevant and impactful. For more mature research directions, it is often hard to understand why things evolved the way they did. I try to provide students with this background and equip them with the technical skills required to be successful in an area. As students mature as scientists, I try to let them be more independent and take charge of their research. ## Harvard Community: 6. **Since starting in the Harvard Statistics Department, what have you valued the most about the Department and Harvard in general?** **Sen:** Overall, it is a privilege to have colleagues who are deeply curious, super motivated, crazy talented, and extremely helpful. This environment has really allowed me to flourish as a researcher and an educator. Two experiences stand out to me: transitioning to teaching online during COVID-19 and adapting to AI in the classroom. I started in 2019 at Harvard and, within a semester, we moved online due to the pandemic! It was our first experience teaching online (and my first teaching position), and my colleagues who had been teaching for decades were in the same position as me. However, everyone was eager to learn and trade strategies on what was working well, how to adapt to the new situation, and how to quickly adopt new tools like Zoom. This same attitude is present in the wider university. I have been particularly struck by how everyone at the University has been adapting to the fast changes due to AI. There are widespread discussions on how to adapt these tools for research and teaching, what is working, and what the pitfalls are. Despite Harvard’s long history, it’s very encouraging that the University has been so forward-thinking on AI integration. See also:- [ Featured Awards ](/news-type/featured-awards)