Mehrtash Babadi, Ph.D.
Director, Computational Methods; Institute Scientist
Mehrtash Babadi is the director of computational methods in the Data Sciences Platform at the Broad Institute of MIT and Harvard, where he is an institute scientist. He leads Cellarium Lab, a research group dedicated to developing open-source production-grade software products and services with strong emphasis on innovation, mathematical rigor, usability, transparency, and reproducibility. Cellarium Lab's primary research topic is developing probabilistic machine learning methods for analyzing single-cell droplet-based and spatial omics data, cellular morphology data, electrophysiology data, and voltage imaging data, with broad application to psychiatric, cardiovascular, and infectious disease research.
Two central themes of Babadi's recent research activities are: (a) precision quantification of raw measurements, enabled by modeling epiphenomena and experimental artifacts on par with biological phenomena; and (b) learning the geometry of massive-scale multi-modal single-cell measurements using self-supervised contrastive and generative AI techniques. His lab actively collaborates with other labs and platforms at the institute, including Ellinor Lab, McCarroll Lab, Macosko Lab, Chen Lab, Blainey Lab, Sabeti Lab, and Farhi Lab. Babadi has also led the Genomic Analysis Toolkit team's effort to develop GATK-gCNV, a germline copy-number variant detection method with clinical-grade accuracy.
Prior to joining the Broad Institute in 2016, Babadi spent two years as a postdoctoral fellow at the Institute for Quantum Information and Matter at California Institute of Technology, where he worked on the theory of non-equilibrium superconductivity and relaxation of far-from-equilibrium quantum many-body systems.
Babadi holds a Ph.D. in theoretical condensed matter physics from Harvard University and B.Sc. degrees in mathematics and physics from Sharif University of Technology.