I am a Professor within the Mullard Space Science Laboratory (MSSL) at University College London (UCL) and lead the Scientific AI (SciAI) research team.

My research interests encompass a wide range of areas across scientific AI, including physics-enhanced AI, geometric AI, statistical AI, generative AI, astrostatistics, Bayesian inference, harmonic analysis, optimisation, and computational techniques. I focus mostly on scientific problems in astrophysics, primarily cosmology, but am also interested in problems in seismology, climate, medical imaging, and computer vision.

I am Founder and CEO of Copernic AI, a startup company developing geometric generative AI techniques. Try out the latest generative AI!

I also offer AI training and consulting services through Neural Learning.

I am founding Director of Research of UCL’s Centre for Data Intensive Science and Industry (DISI) and associated Centre for Doctoral Training (CDT). I was a Core Team member of the ESA Planck satellite mission and a member of the Square Kilometre Array (SKA) Science Data Processor (SDP) working group. I am currently a member of the ESA Euclid satellite Science Consortium and the Rubin Observatory Legacy Survey of Space and Time (LSST) Dark Energy Science Collaboration (DESC) and Informatics and Statistics Science Collaboration (ISSC).

Previously I was a Royal Society Newton Fellow and before that a Leverhulme Early Career Fellow at UCL. Prior to that I was a Scientist at Ecole Polytechnique Federale de Lausanne (EPFL) and a Research Fellow of Clare College, Cambridge, after receiving a PhD from the University of Cambridge.

- Cosmology
- Astrostatistics & Astroinformatics
- Scientific, Physics-Enhanced, Geometric, Statistical & Generative AI
- Bayesian Inference
- Harmonic Analysis

A lightweight python package that implements hybrid sparse-Bayesian dark-matter reconstruction techniques.

Compute the Bayesian evidence (marginal likelihood) from posterior samples generated by any sampling approach.

Reconstruct interferometric observations using learned post-processing and learned unrolled methods.

A lightweight proximal splitting Forward Backward Primal Dual based solver for convex optimization problems.

Compute the Bayesian evidence for high-dimensional log-convex problems by proximal nested sampling.

PURIFY provides functionality to perform radio interferometric imaging, i.e. to recover images from the Fourier measurements taken by …

Scalable Bayesian uncertainty quantification with data-driven (learned) priors for radio interferometric imaging.

S2BALL is a JAX package for computing the scale-discretised wavelet transform on the ball and rotational ball. It leverages autodiff to …

S2FFT is a JAX package for computing Fourier transforms on the sphere and rotation group. It leverages autodiff to provide …

S2LET provides efficient routines for fast wavelet analysis of signals on the sphere. It supports both axisymmetric and directional …

S2SCAT is a Python package for computing scattering covariances on the sphere using JAX. It exploits autodiff to provide differentiable …

S2WAV is a JAX package for computing wavelet transforms on the sphere and rotation group. It leverages autodiff to provide …

The SO3 code provides functionality to perform fast and exact Wigner transforms on the rotation group.

SOPT provides functionality to perform sparse optimisation using state-of-the-art convex optimisation algorithms.

SSHT provides functionality to perform fast and exact spin spherical harmonic transforms based on the sampling theorem on the sphere …