Andrei A. RusuResearch Scientist · Google DeepMindResearcher↗
Researcher
Google DeepMindSenior Research Scientist
Douillard co-developed DiLoCo, a method for training language models across distant computers with far less communication between them.[1] He helped lead Decoupled DiLoCo, which keeps training running through hardware failures, and now works on reinforcement learning to make Gemini more efficient.[2][3]
Google DeepMindOct 2023-present
Google DeepMindAug 2022-2023
HeuritechSep 2018-Jun 2022Work and education history is primarily focused on AI-relevant roles and may not be comprehensive.
The original DiLoCo paper and its authors describe communication-efficient language-model training.
Google credits Douillard among the leads and core contributors to resilient distributed training.
Douillard describes his current work on reinforcement learning for Gemini efficiency.
Based on public sources. If accuracy is important, verify details using the links or references.
For corrections, updates, or requests to remove information, contact us.