Divya Mahajan

Slipstream speeds up recommendation models by skipping unnecessary updates.

A new framework estimates how deep learning models perform on various GPUs.

A new framework improves deep learning training by integrating hardware and task management.