I'm a cosmologist working on large-scale structure, weak gravitational lensing, galaxy clusters, creating redshift catalogues, learning more about intrinsic alignment and many other things. I'm at CosmoStat, CEA Paris-Saclay, and I'm a member of the Euclid consortium, the DESC collaboration and the UNIONS collaboration.
calumhrmurray.github.io · calum.murray@cea.fr
- lectern is a visual editor for the slides an
AI assistant writes. You ask for a talk, it writes the reveal.js file, and you move things
around on a canvas while it carries on with the content. It's on npm as
lectern-editor. - capish does cluster cosmology by likelihood-free inference, written with Constantin Payerne. It forward models a whole cluster survey, from cosmological parameters through to the summary statistics you would actually measure, with the selection and the measurement errors in between.
- This pull request added correlations of arbitrary spin quantities to Arnaud de Mattia's GPU pair counting code. That's what you need to correlate galaxy shapes with galaxy positions, and it runs about a hundred times faster than the CPU codes I was using.
Most of what I write lives in collaboration repositories, or in private ones holding catalogues and results that aren't public yet. Since May 2026 I've been running my analyses by directing agents: I decide the science, the agents write the code. I'm also working with Cail Daley on evaluations for that way of working, on the Lightcone Research stack.