I am a PhD student at Stanford University advised by Percy Liang and Tengyu Ma. My main research interest is in machine learning with domain shifts – can we build machine learning models that perform well even when the train and test distribution are different, or at least signal low confidence when they are not able to make accurate predictions.