MicroCanonicalHMC.jl
Gradient based sampler that uses one single energy level to explore large parameter spaces following Hamiltonian trajectories.
Research Software Engineer
I am a research software engineer at the Advanced Research Computing (ARC) centre of University College London (UCL) developing infrastructure for the LSST and Euclid cosmological surveys to enable new science. I also have a deep interest in accelerating Bayesian inference with gradient methods and Gaussian processes as tools for model-agnostic science.
Click the pictures to head to each repository!
Gradient based sampler that uses one single energy level to explore large parameter spaces following Hamiltonian trajectories.
Fully differentiable Julia code to compute predictions of summary statistics of cosmological observables.
Biggest repository of summary statistics combined in a statistically consistent way in Cosmology using Python