I have a PhD in statistics from Columbia University, and prior to joining IBM was a postdoctoral fellow in the Epidemiology and Biostatistics departments at Harvard.
My main research interests pertain to causal inference, particularly for time-varying treatment strategies. When and how can we infer what would have happened if a population had followed some treatment strategy of interest using data collected under current practice? How can we obtain reliable estimates of optimal treatment strategies? For example, what is the optimal strategy for initiation and dosing of vasopressors in the ICU?