Research
I've spent much of my recent career building models that predict complex scientific systems. But prediction and explanation aren't the same objective. I'm interested in how we build AI systems that can move beyond prediction to discover parsimonious explanations: forming hypotheses, designing discriminating experiments, revising models from evidence, and communicating the resulting understanding to humans. I'm interested in the model architectures and training methods that make those capabilities possible.
I came to machine learning from physics. After a PhD and a spell in academic research, I've spent seven years doing AI research in industry. At Microsoft Research I was a first author on Aurora, a foundation model for the Earth system published in Nature (2025), contributed to Skala, a deep learning model for quantum chemistry, and worked on few-shot learning and molecular representation. I'm now at the Ellison Institute of Technology, building foundation models for biological sequences: proteins, genetics, and whole-cell behaviour.
I enjoy formulating ideas, and I do the technical work to test them myself: designing models, writing code, and running experiments.
The systems I want to build maintain an internal model of their domain and improve it through interaction with data, tools, simulation, and experiment. Biology is a case in point: to be modellable at all it needs both language and verification in the lab. The representations these systems learn about the world fascinate me in their own right. And whatever modelling paradigm we ultimately choose, we should build systems that are transparent, with true explanation accessible to humans.
Experience
Research Scientist
Ellison Institute of Technology · Oxford
- ◆Research lead on foundation models for scientific sequence data, including biological sequences.
- ◆Focused on architecture, pretraining, and scaling, alongside correct evaluation and data curation.
Senior ResearcherJan 2022 – Sep 2025
ResearcherJan 2020 – Dec 2021
Microsoft Research AI for Science · Cambridge & Amsterdam
- ◆First author on Aurora, a large-scale foundation model for the Earth system built on a 3D Swin Transformer backbone with a Perceiver-based encoder, setting a new state of the art for weather forecasting and Earth-system prediction. Published in Nature (2025), used in production by Microsoft weather team and developed by Microsoft AI for Good.
- ◆Member of the Skala team: deep-learning models for quantum chemistry, including learned exchange-correlation functionals (under review, Nature).
- ◆Meta-learning, few-shot learning, and graph neural network research including molecular representation learning; publications at NeurIPS and ICLR.
Data Scientist & Researcher
Faculty Science Ltd · London
- ◆Research team member focused on practical AI safety in customer deployments.
Visiting Scientist
UCSF Department of Radiology, Brain Networks Laboratory · San Francisco
Research Associate (PhD Student)
University of Cambridge · Cambridge
- ◆Thesis: “Photon-mediated entanglement of electron spins in a dynamic solid-state environment”
Selected Publications
Foundation Models & Machine Learning
Aurora: A Foundation Model for the Earth SystemNature 2025
C. Bodnar*, W. P. Bruinsma*, A. Lucic*, M. Stanley*, et al.
Nature 641, 1180–1187 (2025)· *equal contribution
Accurate and scalable exchange-correlation with deep learning
G. Luise et al.
arXiv:2506.14665 (2025, under review at Nature)
Hard Meta-Dataset: Towards Understanding Few-Shot Performance on Difficult TasksICLR 2023
S. Basu, J. Bronskill, M. Stanley, D. Massiceti, S. Feizi.
ICLR (2023)
Fake it until you make it? Generative de novo design and virtual screening of synthesizable molecules
M. Stanley, M. Segler.
Current Opinion in Structural Biology 82 (2023)
Re-evaluating Retrosynthesis Algorithms with SyntheseusNeurIPS 2023
K. Maziarz, A. Tripp, G. Liu, M. Stanley, et al.
NeurIPS AI4Science Workshop (2023) / Faraday Discussions 256, 568–586
FS-Mol: A Few-Shot Learning Dataset of MoleculesNeurIPS 2021
M. Stanley, J. Bronskill, K. Maziarz, H. Misztela, J. Lanini, M. Segler, N. Schneider, M. Brockschmidt.
NeurIPS (2021)
Shapley explainability on the data manifoldICLR 2021
C. Frye, D. de Mijolla, M. Stanley, T. Begley, L. Cowton, I. Feige.
ICLR (2021)
Quantum Physics
Phase-tuned entangled state generation between distant spin qubits
R. Stockill*, M. J. Stanley*, L. Huthmacher*, E. Clarke, M. Hugues, A. J. Miller, C. Matthiesen, C. Le Gall, M. Atatüre.
Phys. Rev. Lett. 119, 010503 (2017)· *equal contribution
Controlling the coherence of a diamond spin qubit through its strain environment
M. J. Stanley et al.
Nature Communications 9, 2012 (2018)
Single-photon emission from single-electron transport in a SAW-driven lateral light-emitting diode
M. J. Stanley et al.
Nature Communications 11, 1–7 (2020)
Full counting statistics of quantum dot resonance fluorescence
C. Matthiesen*, M. J. Stanley*, M. Hugues, E. Clarke, M. Atatüre.
Scientific Reports 4, 4911 (2014)· *equal contribution
Dynamics of a mesoscopic nuclear spin ensemble interacting with an optically driven electron spin
M. J. Stanley, C. Matthiesen, J. Hansom, C. Le Gall, C. H. H. Schulte, E. Clarke, M. Atatüre.
Phys. Rev. B 90, 195305 (2014)
Frequency stabilization of the zero-phonon line of a quantum dot via phonon-assisted active feedback
M. J. Stanley et al.
Applied Physics Letters 105, 172107 (2014)
From the artificial atom to the Kondo–Anderson model: orientation-dependent magnetophotoluminescence of charged excitons in InAs quantum dots
M. J. Stanley et al.
Physical Review B 87, 205308 (2013)
Non-invasive determination of the parameters of strongly coupled 2D Yukawa liquids
T. Ott, M. Stanley, M. Bonitz.
Physics of Plasmas 18, 063701 (2011)
Education
PhD Physics
University of Cambridge · 2017
MSci Physics · First Class Honours
University of Cambridge · 2011
BA Physics · First Class Honours in all years
University of Cambridge · 2010
Awards