David Ross

Engineering Manager at Google Research

David Ross is a Research Manager at Google DeepMind, where he leads the Visual Information and Dynamics team, focusing on computer vision research. His work primarily aims to enhance video understanding, particularly in recognizing objects, motion, and actions. The team has contributed significantly to projects such as the TensorFlow Object Detection API and has developed models that support personal video understanding in Google Photos and Cloud Video Intelligence. They also organize the AVA Challenge to advance spatiotemporal action recognition in videos.1 4

Before joining DeepMind, Ross held various positions at Google, including leading the YouTube Mix team, which developed the personalized algorithmic radio feature for YouTube Music. He has a strong academic background, having earned a Ph.D. in Machine Learning and Computer Vision from the University of Toronto, where his thesis focused on learning probabilistic models for visual motion.1 2

Ross's research has been published in top academic venues, and he has played a pivotal role in several significant projects, including the release of the AVA Actions Dataset, which facilitates research in human activity recognition.1 3

Highlights

Jul 21 · deepmind.google

ICML 2024 - Google DeepMind

Oct 7 · cs.toronto.edu

David Ross - University of Toronto, Canada

Experience

Software Engineering Manager III/L7 at Google (Nov 2016 - Present)

Staff Software Engineer at Google (2012 - Nov 2016), Senior Software Engineer at Google (Mar 2008 - 2012), Intern at Honda Research Institute (2003)

Education

Ph.D. in Machine Learning & Computer Vision from University of Toronto (2004 - 2008), M.Sc. in Machine Learning from University of Toronto (2001 - 2003)