Denis Yarats

Yarats was born in Gomel, USSR, located in modern Belarus. His father worked at a tractor factory and his mother was a housewife. He spent his early life there and in Minsk where he obtained a bachelor's degree in applied mathematics and IT from Belarusian State University. In 2008 he became a finalist in Google Code Jam, but failed to get employed due to poor English skills and preparation. He received his Ph.D. in computer science from New York University, where his research focused on reinforcement learning and natural language.

After his graduation in Belarus, Yarats moved to Moscow where he worked in Yandex maps as a Software Engineer. In his early career, Yarats held research positions at several institutions, including Facebook AI Research (FAIR). His work during this time focused on developing scalable algorithms for learning and decision-making in complex environments.

Yarats worked in Seattle as a Software Development Engineer at Microsoft from 2011 to 2013, where he worked on the Bing search engine. From 2013 to 2016, he worked at Quora as a Staff Machine Learning Engineer in charge of technical projects, and from 2016 to 2022 he worked as an AI Research Scientist at Facebook AI Research (FAIR, now Meta AI).

Yarats co‑authored Image Augmentation Is All You Need: Regularizing Deep Reinforcement Learning from Pixels (Yarats, Kostrikov & Fergus, ICLR 2021), which introduced the DrQ method using simple image-based data augmentations to enable model-free RL algorithms like SAC and DQN to learn directly from pixels and achieve state-of-the-art results on both DeepMind Control Suite and Atari 100k benchmarks.

Yarats AI research has been cited over 9,115 times. His work has been praised by other researchers, including Yann LeCun, who later became an angel investor in Perplexity AI.