Computational Physicist · Scientific Software Engineer

About

I am a Research Software Engineer at the University of Exeter. I build software for simulation and data analysis that researchers run on large data and large compute, and the software stacks beneath it on HPC facilities. Before Exeter, I was the software deployment lead, at Data Management Level 4, for the Simons Observatory’s UK data centre. My Ph.D. is in physics from UC Berkeley, in computational cosmology: I ran the end-to-end analysis behind POLARBEAR’s CMB polarization measurements, and modelled a readout crosstalk systematic for the LiteBIRD satellite, which led me to propose a change in how its detectors are wired.

I am most at home on problems that need the mathematics, the science, the statistics, and the software engineering at once. In each, I aim for software that is correct, feasible at the scale of the data, and reproducible from data to result.

Research Interests

• CMB data analysis and cosmology
• Open quantum system dynamics
• Simulation of physical systems, end to end
• Verification, validation and uncertainty quantification (VVUQ)
• Numerical methods and high-performance computing
• Reproducible environments and workflows
• Agentic engineering

Writing and Talks

• “CMB data analysis in the example of the POLARBEAR and LiteBIRD experiments” — Ph.D. dissertation, also on arXiv
• Statistics
  • “MCMC for a logistic fit from early data” — simulating early data and sampling the posterior, to see how little it says about the limiting value
• HPC and reproducibility
  • HPC guide — best practice for using HPC, independent of any one facility
  • HPC series — an informal series of topics, in small chunks, that will eventually land in the HPC guide
  • “Reproducibility in Computational Research”
• JAX and JIT compilation
  • “A Primer on Just-In-Time (JIT) Compilation”
  • “JIT compilers for scientific computing in Python: Numba vs. JAX” — PyCon UK 2025
  • “Scientific Computing with JAX” — Durham HPC Days 2025
• Agentic engineering
  • “Mathematics is the canary” — what long-horizon coding agents did to mathematicians, and what research software engineers can learn from it
  • “Where do correctness and understanding live in research software?”
  • “Should I use an LLM to refactor my legacy code?”

Connect

• Blog
• GitHub
• Google Scholar
• ORCID
• arXiv
• LinkedIn
• ved.nelok@tcatnoc