My research creates embodied interfaces that enable users to quickly and efficiently adapt generic robots to align with their individual preferences, values, and needs.

I address this problem through two research directions to align robots with users’ preferences: (1) learning user models and robot policies through implicit communication (e.g., interface interactions, engagement, or robot design) and (2) developing data-efficient algorithms that allow users to explicitly customize robot policies. I evaluate my approaches using real robots that interact with real users to address real problems—especially assistive robots for users with limited mobility. I also enjoy building open-source tools (e.g., PyLips and The MUFaSAA Dataset) to support robotics researchers and hobbyists.