Thomas Preud'homme
OpenXLA MLIR Staff Compiler Engineer at Arm
Cambridge, England, United Kingdom
Summary
🤩
Rockstar🎓
Top SchoolThomas Preud'homme is an OpenXLA MLIR staff compiler engineer with 14 years of experience building and optimizing compiler toolchains across GCC, LLVM and now MLIR, primarily in the ARM and ML accelerator space. He has driven low-level, performance-critical work at Arm and Graphcore—from Cortex-A57 FMA steering and endian-aware bswap optimizations to ML-oriented bare-metal LLVM toolchains for IPU tiles. His background combines a PhD in computer science focused on multicore programming with hands-on contributions to bare-metal runtime, binutils, and security fixes like GCC stack-protector hardening. Based in Cambridge, he blends research-grade rigor with pragmatic engineering, often tackling both algorithmic parallelism and architecture-specific code generation. An under-the-radar strength is his track record of improving code size and performance in constrained, hardware-near environments, making him effective at squeezing efficiency from both compilers and silicon.
14 years of coding experience
10 years of employment as a software developer
Bachelor’s Degree, Computational and Applied Mathematics, Bachelor’s Degree, Computational and Applied Mathematics at Université René Descartes (Paris V)Master’s Degree, Computer Science, Master’s Degree, Computer Science at Université Pierre et Marie Curie (Paris VI)Lycée Jean Monnet
French, English, Spanish, Chinese
Github Skills (42)
vulkan10
pytorch10
modular10
python10
data-science10
llvm10
machine-learning10
onnx10
ml10
deep-learning10
tensorflow10
deep-neural-networks10
toolchain10
neural-network10
cuda10
Programming languages (4)
Github contributions (5)
RoboTux/llvm-project
Dec 2023 - Mar 2025
The LLVM Project is a collection of modular and reusable compiler and toolchain technologies.
Contributions:52 pushes, 62 branches in 1 year 3 months
mpercy/pstack
Aug 2011 - Feb 2014
Contributions:23 commits in 2 years 6 months
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.