Andrew Cowie
Andrew Cowie co-authors AI papers on protein structure prediction, reinforcement learning, quantum error correction and AI for mathematics. Acme, a framework for constructing novel reinforcement learning algorithms, appeared on arXiv in June 2020 with him among its authors. Nature published the AlphaFold paper in July 2021, and he was one of its DeepMind authors. In May 2024, Nature published AlphaFold 3, a model that predicts the joint structure of complexes, and he co-authored it. He also co-authored AlphaQubit 2, a neural decoder for topological quantum codes posted in December 2025, and the AI co-mathematician workbench submitted in May 2026.
Timeline
-
June 2020
The Acme reinforcement learning framework
Acme, a framework for constructing novel reinforcement learning algorithms, appears on arXiv in June 2020 with him among its authors.
-
July 2021
A Nature paper on AlphaFold
Nature publishes the AlphaFold paper in July 2021, and he is one of its DeepMind authors.
-
May 2024
AlphaFold 3 for complexes
Nature publishes AlphaFold 3 in May 2024, a model that predicts the joint structure of complexes, with him as a co-author.
-
December 2025
A real-time neural decoder
AlphaQubit 2, a neural decoder for topological quantum codes, is posted to arXiv in December 2025 with him as an author.
-
May 2026
The AI co-mathematician workbench
Submitted to arXiv in May 2026, the AI co-mathematician paper describes a workbench for mathematicians, and he is one of its authors.
Sources and credits
Photo: We have no published photo of Andrew Cowie yet. Send us a free-licensed photo.
- [2006.00979] Acme: A Research Framework for Distributed Reinforcement Learning
- Highly accurate protein structure prediction with AlphaFold | Nature
- Accurate structure prediction of biomolecular interactions with AlphaFold 3 | Nature
- Accurate structure prediction of biomolecular interactions with AlphaFold 3 - PMC
- [2512.07737] A scalable and real-time neural decoder for topological quantum codes
- [2605.06651] AI co-mathematician: Accelerating mathematicians with agentic AI