Eric Oermann, MD

“I’m an associate professor of data science and neurosurgeon at NYU Langone. In many ways, my career’s split between two hats: my clinical hat and my research hat. My lab is located within the Medical Center and the fact that we’re physically located in the hospital, rather than being in the Center for Data Science downtown, really captures the mission of the lab. Our major focus is: how do we use AI to actually impact clinical care?

“For my residency, I lucked out in matching at Mount Sinai, uptown. Dr. Bederson, my Chairman there, is a phenomenal skull-based surgeon, and very into thinking outside the box. Sort of like Ewend, he told me, ‘Software engineering is a super cool outside interest. I don’t know how you’re going to integrate it with medicine, but I’m going to do everything in my power to facilitate that.’ It led to all these amazing opportunities. I did research that got me into Forbes Magazine, I got to work at Google as a software engineer, and I did a start-up where we knocked together these machines and started prototyping on them—I think it was one of the first real AI research projects physically located within a hospital.

“At NYU, the Dean has given us a phenomenal amount of computing. We have arguably the best applied math department in the world over at NYU Courant, and the Center for Data Science was founded by Yann LeCun, one of the fathers of modern AI. If you want to apply AI to healthcare, I can’t think of a better environment for it.

“The cool thing about being an AI research group is that you’re naturally interdisciplinary. There’s neurology, neurosurgery, radiation oncology, oncology, the Center for Data Science—I mean, ten or so departments—working across our projects. One of our big ones right now is working with large language models. We’ve been very interested in building these models to understand medical language—to actually build them against specific operational and clinical problems that we have in the hospital, and deploy them live for clinical effect.

“I think everyone’s betting on AI, but we’ve still seen fairly marginal real-world impact. What we’ll hopefully start seeing is a shift toward what’s called the ‘last mile’ problem. We do all this work to get AI technologies validated on healthcare data and published on healthcare problems, but they don’t actually get deployed to impact patient care. It’s that last-mile impact I’m hoping we’ll see in the next ten years. “I spent a year at Verily Life Sciences at Google. The premise was, ‘let’s try to disrupt healthcare by hiring a bunch of really smart doctors and software engineers, putting them together in teams and seeing what comes out of that.’ I try to run my lab like that; we physically connect physicians with computer scientists.”

As told to Parker Menzimer.