Summary
Laurent Montigny is a system software engineer with nine years of deep expertise accelerating AI and HPC workloads across CPU and GPU architectures, currently optimizing large-scale LLM inference on GPU clusters. He has led a 20-person team at Intel driving an LLVM/MLIR-based AI graph compiler and delivered cross-stack performance wins on Intel Data Center GPUs and Xeon CPUs. His background spans AMD ROCm optimization, supercomputer architecture prototyping, and CUDA/C++ kernel development for CFD and simulation, giving him a rare blend of compiler, runtime, and low-level hardware experience. Laurent pairs hands-on coding (C++, Python, CUDA) with system-level performance analysis and mentoring, having also supported startups through Intel Ignite. An interesting through-line: he started in fluid mechanics and CFD research, coupling reinforcement learning with MPI-parallelized simulations, which informs his pragmatic approach to large-scale, physics-informed compute problems.
8 years of coding experience
9 years of employment as a software developer
Master's degree, Fluid Mechanics, Master's degree, Fluid Mechanics at École Polytechnique
Master thesis, Computer Science, Master thesis, Computer Science at ETH Zürich
Master’s Degree, Fluid Mechanics, Master’s Degree, Fluid Mechanics at Pierre and Marie Curie University
English, German, French