Core contributor to the Gemini 2.5 technical report (2025).
Published multiple papers at top-tier conferences: CVPR (including Highlight papers Super-CLEVR and Causal-CoG), ICCV, EMNLP, NeurIPS, EACL, NAACL, etc.
Ph.D. thesis titled 'On the Diagnosis and Generalization of Compositional Visual Reasoning' (2024).
Released code and datasets for several projects, including Super-CLEVR, ExoViP, and 3D-Aware VQA.
Invited talk at the Computational Cognitive Science Lab, MIT (May 2023).