I build vision-language and robot learning systems that unite perception, reasoning, and action. Work spans open-vocabulary perception, text-to-motion generation, simulation-based post-training, human-object interaction, and multi-task optimization for data-efficient, generalizable skills.