Google Research

Keren Gu


Before joining the residency program, I spent three years at Sift Science -- a startup in San Francisco using machine learning to stop online abuse -- as a machine learning engineer launching ML powered products and experiencing the startup hypergrowth first hand. Prior to that, I double majored in mathematics and computer science from MIT and completed my Master's with Professor Julie Shah on interactive robotics at CSAIL.

Being an AI resident at Google for the past few months has been exhilarating! With no prior deep learning experience, I've already learned so much about computer vision and TensorFlow. I am currently working with amazing mentors on fundamental deep learning research -- understanding the generalization power of existing regularization methods and distortion techniques for image classification beyond ImageNet test set. We are hoping to shed light on questions such as "Are today's state of the art (SOTA) ImageNet models overfitting to the test set?" and "How would today's SOTA models fare in the real world?"

Looking ahead, I'm excited to be exploring other areas of AI during my year. I am planning my next project on deep learning applications to the medical domain as well as another 20% project dabbling in AI policy. I also plan to explore Google offices around the world, starting with London and Amsterdam!

Research Areas

Learn more about how we do research

We maintain a portfolio of research projects, providing individuals and teams the freedom to emphasize specific types of work