Ang Qi

I’m Ang Qi, an undergraduate at the University of Southern California majoring in Applied and Computational Mathematics with a minor in Computer Science. I expect to graduate in May 2027.

I am a research assistant in USC’s Physical Superintelligence Lab, where I work on the robustness and reliability of vision-language models. My work focuses on understanding model failures, building multimodal data and evaluation pipelines, and investigating targeted model repair through post-training and parameter-efficient fine-tuning.

My research spans visual reasoning tasks such as object recognition, attributes, spatial relationships, counting, and negation. I am interested in how models generalize after targeted repair and how we can make their behavior more reliable.

Explore my research, publications, and CV.