Reuben Tan

Reuben Tan is a researcher at Microsoft Research in Redmond, where he works in the Deep Learning Group under Dr. Jianfeng Gao on agentic models for multimodal reasoning. He completed his PhD at Boston University under the advisement of Professors Kate Saenko and Bryan Plummer, with close collaboration with Dr. Bryan Russell at Adobe Research. His research spans multimodal learning and video understanding, and he is known for contributions including the Magma foundation model for multimodal AI agents, the Koala framework for long video understanding, and the LAPA method for unsupervised pretraining of vision-language-action models.