Jean-luc Duprat
Member Of Technical Staff at OpenAI
San Francisco, California, United States
Summary
👤
Senior🎓
Top SchoolJean-luc Duprat is a senior staff engineer and technical leader with over two decades of experience building compilers, GPU backends, and real-time graphics systems for top-tier hardware and media companies. He has led and scaled engineering teams at Apple and Meta, architecting high-performance code generation and accelerator enablement for cutting-edge GPU architectures that power consumer devices. His background spans chip microarchitecture, concurrency and vectorization, geometric modeling, and shader/renderer development, with hands-on proficiency in C/C++, assembly, Python and performance optimization. A pragmatic problem-solver, he has contributed to prominent open-source efforts like ISPC, improving cross-version Python build tooling and testing for SIMD compilers. Combining deep math and physics training with production delivery across startups and large enterprises, he excels at turning complex hardware and compiler research into robust shipping systems.
14 years of coding experience
26 years of employment as a software developer
PhD (Not completed) Computer Graphics, PhD (Not completed) Computer Graphics at University of VictoriaM.Sc. Computer Science, M.Sc. Computer Science at The University of British ColumbiaÉcole Européenne Bruxelles IIB.Sc. Physics, B.Sc. Physics at McGill University
French, English
Github Skills (12)
simd10
compiler10
programming-language10
compiler-compiler10
intel10
build-automation9
performance-analysis8
performance-monitor8
performance-analytics8
performance-tuning8
performance-measurement8
performance-monitoring8
Programming languages (4)
Github contributions (5)
ispc/ispc
Jan 2012 - Jul 2013
Intel® Implicit SPMD Program Compiler
Role in this project:
Back-end Developer
Back-end DeveloperContributions:83 commits in 1 year 6 months
Contributions summary:Jean-luc contributed to the Intel® Implicit SPMD Program Compiler (ISPC) project by modifying Python scripts used in the build process. These changes focused on ensuring compatibility with both Python 2.x and 3.x releases, and fixing a few issues related to memory access and the correct handling of strings in the build environment. The user also improved testing infrastructure and added KNC support.
compilerspmdispcprogramming-languageintel
jduprat/flash-attention
Jul 2025 - Mar 2026
Fast and memory-efficient exact attention
Contributions:6 pushes, 2 branches in 8 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.