I am a Career Development Research Fellow at St John’s College, Oxford, and a Hintze Prize Fellow in Cosmology at the University of Oxford. I study how structure formed in the nearby Universe, out to a few hundred megaparsecs (roughly a billion light-years), and how fast the Universe is expanding.
I specialise in Bayesian inference for noisy measurements and incomplete samples, including datasets with complex calibrations and systematic uncertainties. I develop statistical models and inference software in Python and JAX, using GPUs and high-performance computing clusters to run large simulations and sample models with millions of parameters. I am interested in using agentic systems to accelerate this method development, shortening the path from an idea to an inference pipeline and shifting the effort onto the physical modelling and results interpretation.
I also work on simulation-based inference and differentiable physical models, and use machine learning to emulate expensive, high-dimensional simulations. See Methods for the statistical and computational side of my work, and Research for the cosmology.
I am a member of the Aquila Consortium, Learning the Universe, and CAMELS collaborations.
Brief background
I completed my DPhil in Astrophysics at the University of Oxford (Balliol College), where I was a Snell Exhibitioner, supervised by Harry Desmond, Julien Devriendt, and Adrianne Slyz. Before Oxford, I completed my MSc in Physics at the Ludwig Maximilian University of Munich, with my thesis on strong-field lensing of gravitational waves externally supervised by Miguel Zumalacárregui and Marius Oancea, and my BSc in Physics with Astrophysics at the University of Glasgow, which included a year-long exchange at the Hong Kong University of Science and Technology. At Glasgow, I worked with John Veitch and Chris Messenger on gravitational-wave data analysis.
While at Oxford, I held a Pre-Doctoral Fellowship at the Center for Computational Astrophysics, Flatiron Institute in New York, working with Shy Genel and Lucia A. Perez, and was a visiting researcher at the Institut d’Astrophysique de Paris working with Guilhem Lavaux.
I am always happy to talk about research, collaborations, or science in general. Beyond cosmology, I advise research groups and organisations on Bayesian modelling and inference for noisy or incomplete data, and on accelerating inference for expensive models with GPUs, emulation, and gradient-based sampling.
For research and academic enquiries, email me at richard.stiskalek@physics.ox.ac.uk, otherwise richard.stiskalek@protonmail.com.
