Debdeep Pati

The Department of Mathematics & Statistics at Boston University is seeking a postdoctoral researcher to work with Professor Debdeep Pati on probabilistic modeling and approximate Bayesian inference . The project centers on building scalable, theoretically grounded methods for inference in complex scientific settings, combining statistical methodology, mathematical analysis, and computational implementation. The research spans three connected areas: Foundations of approximate Bayesian inference , including variational inference, expectation propagation, and related approximation algorithms, with attention to convergence, approximation accuracy, and uncertainty quantification. Probabilistic inference for generative and simulation-based models , especially high-dimensional latent-variable models, diffusion-based generative models, and foundation models where likelihood evaluation is difficult or simulations are expensive. Scientific applications and inverse problems , developing robust inference and uncertainty quantification tools for physics-based simulations and high-dimensional scientific observations. The successful candidate will work closely with Professor Pati in Boston University’s Department of Mathematics & Statistics and will have opportunities to develop independent research directions within these themes. The position is research-focused and designed for candidates interested in combining modern Bayesian computation with rigorous theory and applications to scientific modeling. The appointment is expected to begin in Summer or Fall 2027.