My current research is at the intersection of NLP, HCI, and Psychometrics, where I design NLP tasks and conduct human studies to evaluate model performance, from which we aim to teach the model to learn like how a human learns. I also design interactive human-in-the-loop ML tools. My work has been published in HCI venues such as CSCW, CHI and NLP conferences such as ACL and NAACL.