Amir Zamir

Amir Zamir is a Tenure-Track Assistant Professor of Computer Science at the Swiss Federal Institute of Technology (EPFL) and leads the Visual Intelligence and Learning Lab (VILAB). Previously affiliated with UC Berkeley, Stanford, and UCF, his research spans computer vision, machine learning, and AI, focusing on embodied intelligence and multimodal learning.

• Current Appointments: EPFL (Tenure-Track Assistant Professor)
• Industry Experience: Chief Scientist at Aurora Solar (2015-2022), Chief Scientist at Duranta

His research emphasizes advancing vision systems beyond narrow approaches toward general multi-modal/multi-task models that operate in real-world environments. Key projects include 4M (Multimodal Foundation Models), Taskonomy (Task Transfer), Gibson (Sim-to-Real), and Omnidata (Steerable Datasets). Current work explores flexible-length tokenization (FlexTok), visual personalization (ViPer), and robust cross-domain learning.

Recent publications demonstrate expertise in autoregressive image generation, vision-language modeling, and embodied robotics. Awards include the ECCV Young Researcher Award, NVIDIA Pioneering Research Award, and multiple CVPR/SIGGRAPH best paper recognitions. He advises PhD students in vision systems, AI, and robotics, and teaches courses on Visual Intelligence, AI Product Management, and Autonomous Robotics.