About Me
I am the Director of Clinical AI at Rajpurkar Lab (Department of Biomedical Informatics, Harvard Medical School), where I lead research on building and evaluating multimodal AI systems for medicine. Our ultimate goal is to develop and validate AI systems that can integrate visual, textual, audio and other data to perform at a specialist physician's level and communicate in natural language (see our review on multimodal AI in Nature Medicine and our perspective on voice agents in NPJ Digital Medicine). In parallel, I am a part-time Scientist at a2z Radiology AI, where I lead the clinical validation of AI models and regulatory interactions with the FDA and EU regulators.
Previously, I was a research scientist at RadAI, where I trained and deployed radiology AI models now used by hundreds of radiologists nationwide. Before that, my postdoctoral work at Yale University focused on leveraging large-scale datasets (EHR, genomics, and neuroimaging) to investigate cerebrovascular disease and brain health. This research established novel links between blood pressure and white matter integrity, explored the influence of genomics on brain health, and shed light on cardiovascular health disparities. A key focus of my work involved using Mendelian randomization to identify causal risk factors of stroke, such as the influence of smoking on subarachnoid hemorrhage, and the link between LDL cholesterol and both intracerebral hemorrhage and subarachnoid hemorrhage. I summarized the potential of Mendelian Randomization in a review.
My journey began with a medical degree (magna cum laude) from the National University of the Northeast in Argentina, followed by a neurology residency at Fleni, the leading neurological institute in Argentina.
Recent Publications
Full list on Google Scholar.
Hobbies
Beyond my professional work, I enjoy a mix of activities:
-
Tennis & Pickleball: USTA social leagues and pickeball weekend open play.
-
CrossFit: Morning WODs at JP CrossFit.
-
Sci-Fi: Diving into sci-fi literature and film.
