Michael Figurnov
Senior Staff Research Scientist · Google DeepMind
I work on pretraining of Gemini models.
Previously, I was a core machine learning researcher behind AlphaFold 2, with contributions to its architecture, training, and predictive accuracy. My earlier research focused on probabilistic machine learning, gradient estimation, and adaptive computation.
Research
Full CV2026–present
Gemini pretraining
2022–2025
AlphaFold 3 & AI for biology
Co-author of AlphaFold 3. Research on protein modelling and LLMs for biology.
2019–2021
AlphaFold 2
I designed the Evoformer architecture, co-developed the recycling mechanism, and improved model training and predictive accuracy. I also trained the models used in AlphaFold 2’s winning CASP14 submission.
My architectural contribution was highlighted in John Jumper’s Nobel lecture.
Career & education
2018–present
Google DeepMind, London
- Senior Staff Research Scientist2024–present
- Staff Research Scientist2021–2024
- Senior Research Scientist2019–2021
- Research Scientist2018–2019
2013–2019
PhD in Computer Science
Higher School of Economics, Bayesian Methods Research Group, supervised by Dmitry Vetrov. Dissertation (in Russian).
2014–2016
Internships & research visits
Microsoft Research Cambridge (2014), University of Toronto (2015), Google Seattle (2016).
2008–2013
Applied Mathematics & Computer Science
Specialist degree, Moscow State University.
Academic service
Reviewer for JMLR, TMLR, NeurIPS, ICML, CVPR, the Symposium on Advances in Approximate Bayesian Inference (AABI), and the Machine Learning for Structural Biology (MLSB) workshop.
Updates
Started working on pretraining of Gemini models at Google DeepMind.
My work on the Evoformer architecture was highlighted in John Jumper’s Nobel lecture.
AlphaFold was recognised with the 2024 Nobel Prize in Chemistry (Demis Hassabis and John Jumper).
Promoted to Senior Staff Research Scientist.
Released AlphaFold 3.
Published the Nature paper describing AlphaFold and released the source code and a database of predicted structures.
Promoted to Staff Research Scientist.
Announced AlphaFold 2 at CASP14.
Defended my PhD in Computer Science at the Higher School of Economics.
Joined DeepMind as a Research Scientist.
Selected publications
All publications
* Equal contribution.
Talks & teaching
Bayesian Methods Research Group Seminar, 2021
