About

Benjamin Thorne

Lead Engineer at Atomic Industries, San Francisco

I build optimization and machine-learning systems grounded in physics. At Atomic I work on software that starts from the geometry of a part and designs the injection mold that makes it. Before that I was a cosmologist. I worked on the microwave sky at Oxford, Princeton, Tokyo, Davis, and Berkeley.

Email LinkedIn GitHub Scholar CV (PDF)

What I work on

Injection molding makes most of the plastic parts around you, and every one of them needs a mold designed by hand. There are few, if any, established techniques for automating tool design at industrial scale. Atomic is building them. My team works on the design software: the stages that turn the geometry of a part into the geometry of its mold, from parting the part to laying out cooling and ejection.

The work combines computational geometry, physical simulation, and large-scale optimization, with machine learning where a learned model beats a direct computation.

Selected work

Automated tooling designAtomic Industries · 2023–

The platform that takes customer CAD through to manufacturable injection-mold tooling. Computational geometry, GPU compute, and large-scale optimization on a hard physical design problem. We use Julia, Python, Rust, and C++, and we run on hardware we own.

Differentiable cosmologyUC Davis · Berkeley · 2020–2022

A differentiable foreground extension to CMBLensing.jl, accelerated with CUDA.jl. Sparse approximations and preconditioners cut the cost of a log-likelihood evaluation by roughly 100×, which made inference across tens of A100 nodes practical. Paper

FAIR UniverseBerkeley Lab · NeurIPS 2024

A ten-plus person collaboration across Berkeley Lab, the University of Washington, Paris-Saclay, and ChaLearn that built uncertainty-aware machine-learning benchmarks for high-energy physics, served on the Perlmutter supercomputer. Paper · Project

Python Sky ModelOpen source · 2017–

The de facto standard simulation package for the microwave sky, with 300+ citations. Used by the Simons Observatory, CMB-S4, and the Atacama Cosmology Telescope; now maintained by the Pan-Experiment Galactic Science group. GitHub · Paper

Background

2023–now
Lead Engineer, Atomic Industries, San Francisco · previously Senior MLE
2023
Machine Learning Engineer, Lawrence Berkeley National Laboratory (NERSC)
2019–2022
Postdoctoral Researcher, University of California, Davis
2015–2019
PhD in Astrophysics, University of Oxford, with Princeton University and Kavli IPMU
2011–2015
M.S. & B.S. in Physics, University of Oxford, New College · first-class honors

Writing

All publications →