Ian Williamson
Staff Research Engineer at Google DeepMind
San Francisco, California, United States
Summary
👤
Senior🎓
Top SchoolIan Williamson is a Staff Research Engineer at Google DeepMind with 11 years of experience designing and implementing ML-driven solutions for computational nanophotonics and inverse design. He led AI-guided photonic design efforts at X, translating research-stage optimization and differentiable simulation techniques into scalable engineering workflows. His background spans a PhD in electrical engineering, postdoctoral work at Stanford, and hands-on development of differentiable physics tooling—most notably a JAX-Meep wrapper that enables gradient-based optimization of FDTD electromagnetic simulations. Based in San Francisco, he combines deep domain expertise in photonics with practical software engineering and systems thinking, often bridging research prototypes and production-ready code. An understated strength is his track record of integrating adjoint methods into modern ML stacks to make complex simulation-driven design both differentiable and maintainable.
11 years of coding experience
9 years of employment as a software developer
PhD Electrical Engineering, PhD Electrical Engineering at The University of Texas at Austin
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (9)
simulations10
simulation10
adx10
adp10
jax10
python10
adt10
machine-learning9
optimization7
Programming languages (7)
Github contributions (5)
NanoComp/meep
May 2021 - Sep 2021
free finite-difference time-domain (FDTD) software for electromagnetic simulations
Role in this project:
ML Engineer
ML EngineerContributions:29 reviews, 7 commits, 7 PRs in 4 months
Contributions summary:Ian primarily contributed to the development of a JAX-Meep wrapper, enabling the differentiation of Meep simulations using JAX. Their work involved the initial implementation, testing, and refinement of the wrapper, including the integration of adjoint modules for gradient calculations. They also addressed import issues, refactored code for clarity, and applied code formatting to improve readability and maintainability. This suggests a focus on integrating differentiable programming capabilities within the Meep simulation framework for potential use in optimization or inverse design problems.
electromagneticfdtdfinite-differencesimulationssimulation
fancompute/FDFD.jl
Mar 2018 - Jan 2021
Pure Julia implementation of the finite difference frequency domain (FDFD) method for electromagnetics
Contributions:117 commits, 1 PR, 67 pushes in 2 years 10 months
frequency-domainphysicssymmetryelectromagneticsfenics
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.