I am an Associate Professor (Reader) in Statistical Machine Learning in the Department of Mathematics at Imperial College London. My research develops principled, efficient methods for generative modelling and decision-making under uncertainty—grounded in probability, optimisation, and numerical analysis—and applies them to complex physical and engineered systems.
I serve as Editor-in-Chief of ACM Transactions on AI for Science (TAIS) and as an Associate Editor for the SIAM/ASA Journal on Uncertainty Quantification.
Research
I aim to make generative AI trustworthy, compute-efficient, and scientifically useful: models that reason with uncertainty, honour constraints from physics and engineering, and deliver high performance under realistic compute and data budgets. Current themes include:
- Generative modelling: diffusion and flow-matching models, energy-based models, and lean LLMs; structure- and physics-informed training and sampling.
- Efficient inference & sampling: test-time scaling and inference-time optimisation for LLMs and other generative models.
- Scientific automation: workflows that accelerate hypothesis generation, design, and model-based experimentation.
- Decision-making under uncertainty with strict compute budgets: robust control and sequential decision-making for simulation-heavy systems.
Applications include biochemical systems, chemical engineering, and the monitoring and control of complex physical and engineered systems across defence, aerospace, supply chains and logistics, and energy systems.
Publications
See my Google Scholar for an up-to-date list.
Previous Roles
- Director of Fundamental Research in AI — The Alan Turing Institute
- Senior Research Scientist & Principal Scientist — Improbable Defence & National Security
- Group Leader, Data-Centric Engineering Programme — The Alan Turing Institute
- Lecturer in Probability & Statistics — University of Sussex
Education
- PhD, Mathematics — University of Warwick
- MSc, Scientific Computing — University of Warwick
- BSc (Hons), Mathematics & Computer Science — University of Malta
Research Group
Current PhD Students
- Rafael Athanasiades (StatML CDT, Year 1)
- Edoardo Monti (Marie-Skłodowska Curie DN BLESSED)
- Paula Cordero Encinar (StatML CDT, co-supervised with Deniz Akyildiz)
- Paul Valsecchi Oliva (StatML CDT, primary supervisor: Deniz Akyildiz)
- Tobias Schroeder (Final year, co-supervised with Greg Pavliotis)
Former PhD Students
- Enrico Crovini (co-supervised & co-funded by UKAEA)
- Xing Liu (primary supervisor: Axel Gandy)
- Yanni Papandreou (co-supervised with Jon Cockayne, Southampton)
- George Wynne
- Nik Nüsken (primary supervisor: Greg Pavliotis)
Former Postdocs
Prospective Students & Collaborators
I welcome enquiries from motivated PhD candidates and collaborators with backgrounds in statistics, applied mathematics, machine learning, or control/optimisation. Please include a CV, a short statement of interests, and links to any relevant code or publications.
Links
Contact
Email: asurname@imperial.ac.uk