My google scholar page | arXiv
My research lies at the interface of applied probability, computational methods, and experimentation. My current research is inspired by ideas in algorithmic fairness, reinforcement learning and sequential decision making and their connections to Markov Chain Monte Carlo (MCMC) methods. I also have an active interest in using ML/statistical modeling to address problems in climate science. Below is a selected bibliography.
I study Markov chains both in discrete and continuous state spaces. For discrete state space Markov chains, I am interested in studying their mixing times often via elegant coupling techniques. In continuous state spaces, mixing times do not always give meaningful answers. Instead, I use scaling methods and diffusion limits to obtain practical guidelines for implementing them.
I am interested in developing statistical methods with a solid theoretical backing. This philosophy has often led me to unexpected discoveries and finding connections between disparate areas of research. I maintain an active interest in Bayesian methodology.
I was exposed to causal methods and experimentation at Harvard. In recent years, especially because of my work in the industry, I am drawn to research problems that require application of causal methods in new application domains.
I am an affiliate member of HUCE and keenly follow various climate related research at Harvard and beyond. I am interested in addressing issues due to poor data quality, synthesizing different sources of data, spatial modeling and relating climate and health.