Hi! I am a Ph.D. candidate at the Univeristy of British Columbia (UBC) and the Vector Institute for AI advised by Vered Shwartz and Raymond NG in the Natural Language Processing group. I frequently collaborate with Leonid Sigal in the Computer Vision group.

I work at the intersection of NLP and Computer Vision. Specifically, I evaluate and improve reasoning capabilities across multiple modalities (text, images, videos) - which are fundamental for safe and effective deployment in real-world applications such as embodied agents and AR/VR. Drawing inspiration from human cognition, I have worked on fundamental reasoning abilities including causal, counterfactual, and commonsense reasoning, as well as cultural and social norm understanding.

Currently, I am investigating the robustness of Video-LLMs to unexpected scenarios. My recent work Black Swan evaluates models on how they explain and adapt in unpredictable video events. In my ongoing work SPIKE, I am developing post-training methods inspired by Bayesian Theory of Mind that enable VLMs to revise their beliefs and become more resilient to novel scenarios.