What Is arthurdouillard.com?
Updated
arthurdouillard.com is the personal academic website of Arthur Douillard, a researcher at DeepMind in London. The site is a single-page professional profile: it states his current research focus, summarizes his background, and links to three sections — Research, Tools, and DeepCourse. It is not a company site, a product page, or a blog with dated posts; it is a personal landing page for his work.
Who the site belongs to
The page identifies Arthur Douillard as a researcher at DeepMind in London. According to the site's own description, his current work is on reinforcement learning aimed at scaling intelligence per FLOP for Gemini. Before that, he worked on AI for systems, specifically distributed and resilient pre-training. He also completed a PhD in continual learning for computer vision.
That combination covers three distinct areas, which is useful to keep separate when reading the site:
- Reinforcement learning — his stated current focus, tied to scaling compute efficiency for Gemini.
- Distributed resilient pre-training — prior work on AI systems, i.e., training large models across many machines and keeping that training robust.
- Continual learning for computer vision — his PhD topic, concerned with models that keep learning new tasks without forgetting earlier ones.
What the site contains
The page is organized around three entry points:
| Section | What it points to |
|---|---|
| Research | His research work and publications |
| Tools | Software or utilities he has released |
| DeepCourse | A course-related resource associated with the site |
The visible page content is essentially a short biography plus these three links. There is no pricing information, no login requirement mentioned, and no indication of a paid product — the site reads as an academic and professional profile rather than a commercial offering.
When this site is useful to you
- You want to know who Arthur Douillard is and what he works on before citing, contacting, or following his research.
- You are looking for his publications or code, and want the canonical starting point rather than a third-party mirror.
- You are interested in continual learning, distributed pre-training, or RL for large-scale models and want to see how one researcher frames those areas.
- You came across "DeepCourse" or one of his tools and want to confirm it originates from him.
If you need a company's product documentation, pricing, or support, this is not that kind of site. If you need a researcher's own summary of his work and links to his output, it is the right place to start.