Siavash Raissi | Master of Medical Sciences in Biomedical Informatics

When Siavash Raissi talks about his work, he sounds like someone who has already lived a few different academic lives. Now a second-year student in Harvard Medical School’s Master of Medical Sciences in Biomedical Informatics program, he is learning to bridge disciplines and prepare for a career as a physician-scientist who can translate computational insights into better care at the bedside.

Raissi graduated from Tufts University in 2024 with a BS in biology and minors in computer science and art history. In Peter Bullock’s lab at Tufts, Raissi applied protein design techniques to evaluate the efficacy of immune checkpoint inhibitor antibodies, an experience that sharpened his interest in translational science.

While his peers prepared to go straight to medical school after completing their undergraduate degrees, Raissi wanted to explore his computational interests more deeply before committing to a strictly clinical path. “I’ve always been interested in computer science, and I figured that there were a lot of really interesting ways that I could pair it with biology in the realm of computational biology or biomedical informatics,” he explains.

At a computational biology seminar in Boston, Raissi discovered Harvard’s MMSc in Biomedical Informatics program somewhat serendipitously. “I sat next to someone who was in the program,” he recalls. “They told me how much they enjoyed it, how it let them merge their backgrounds in biology and computation. Their story sounded a lot like mine. That conversation made me look it up and apply.”

For Raissi, the chance to do a full year of thesis research was the deciding factor. “The thesis component was really what attracted me,” he says. “It’s rare in a master’s program to have protected time to do a full year of research. That was the opportunity I was looking for.”

Today, Raissi works in the lab of Kun-Hsing Yu, MD, PhD, in the Department of Biomedical Informatics. Yu’s group is known for pushing the boundaries of how AI can extract insights from pathology images. Raissi’s thesis research focuses on improving the interpretability of pathology AI models. These tools have shown impressive abilities in areas like cancer classification and genotype inference, but their use in the clinic is often limited by a lack of explainability.

“My project focuses on developing an accompaniment or an additional component of these models that can help interpret their outputs and relate them back to nuclear morphology,” he says. The problem is technical, but the implications are human. If AI tools can show clinicians how they make decisions, adoption in clinical practice becomes far more likely. “Ultimately, it’s about taking something powerful in the lab and making sure it can actually help patients.”

Raissi admits that his background didn’t fully prepare him for the world of data science. “Even though I minored in computer science, I had never done real data science,” he explains. “And that’s a different skill set. This program was my first deep dive into it.”

That made the two-year format of the program critical. The coursework, mentorship, and—most importantly—the protected research time gave Raissi space not only to take on an ambitious thesis project but also to develop the data science skills he lacked coming in. During his time in the program, he has been able to move from a newcomer in the field to someone who feels confident working within it.

Raissi credits his advisor, Dr. Yu, for much of that growth. “I meet with him once every two weeks. I also work with a postdoc that I see regularly, Dr. Katharina Hoebel. Both of them are always willing to give me advice on my project, but they also encourage me to pursue that knowledge and this project on my own as well.”

Collaboration extends beyond the lab. Raissi describes his cohort as tight-knit, with projects often spilling beyond the classroom. “In Dr. Marinka Zitnik’s course, I worked with two of my fellow master’s students. We turned our class project into a full paper, and have begun submitting it to conferences and presenting it in different areas.” The project, RADGame, is an online education platform for training radiologists using AI-powered feedback. “That kind of thing is only possible because the program encourages us to be ambitious and work together.”

One of the unexpected benefits of the program has been the range of connections that Raissi has been able to build beyond Harvard Medical School. Cross-enrollment opportunities at other Harvard schools and MIT, and access to Harvard’s affiliate hospitals, such as Dana Farber Cancer Institute, have opened doors to collaborations and mentorships he had not foreseen. That breadth reinforced his view of biomedical informatics as a discipline defined by bridges between fields, institutions, and, ultimately, between research and the clinic.

As he applies to medical school, Raissi is committed to becoming a physician-scientist, though he hasn’t yet chosen a specialty. Pathology, closely tied to his current research, remains a strong possibility, but he is also considering nephrology, radiology, and even primary care. For now, he is keeping an open mind as he thinks about how to balance clinical practice with research.

What has crystallized, though, is his commitment to research. When Raissi entered the program, he was uncertain about how central research would be to his future career. Over time, exposure to different subfields of biomedical informatics helped him develop a strong passion for the work and the confidence to see it as an essential counterpart to medicine.

For prospective students, Raissi has clear advice: “I’ve been able to thrive here. What matter most are curiosity and interest. Explore the faculty websites, see what excites you, and if you can picture yourself pursuing those projects, apply. This program will prepare you to succeed.”

Written by Bailey Merlin