Senior AI Inference Systems Engineer – GPU-Optimized ML

NVIDIA; France; Hybride; EUR 120 000 - 190 000; Plein temps

Résumé du poste

NVIDIA in France is seeking a senior ML systems engineer to advance large-scale AI inference on GPU hardware. You will contribute features to vLLM, optimize inference pipelines, and help push the pareto frontier of performance.

You’ll design and benchmark GPU kernels, build high-level DSLs, and orchestrate containerized deployments on GPU clusters across clouds. Strong Python/C++ skills and deep knowledge of ML systems are required.

Qualifications

• Bachelor’s degree (or equivalent) in CS/CE/SE with 7+ years, or Master with 5+ years, or PhD in ML Systems/GPUs.
• Strong Python and C/C++ programming; Go or Rust a plus; solid CS fundamentals.
• Knowledgeable about performance engineering in ML frameworks and inference engines.

Responsabilités

• Contribute features to vLLM and optimize inference with new NVIDIA GPU features.
• Develop, optimize, and benchmark GPU kernels with fusion, autotuning, and memory/layout optimizations.
• Define and build inference benchmarking methodologies; contribute MLPerf submissions.
• Architect scheduling and orchestration of containerized large-scale inference on GPU clusters across clouds.
• Publish original research pushing the pareto frontier in ML Systems and integrate ideas into NVIDIA software.

Connaissances

Python

C/C++

Go

Rust

Algorithms

Operating Systems

Computer Architecture

Parallel Programming

Distributed Systems

DL Theories

Formation

Bachelor’s degree in CS/CE/SE

Master’s degree in CS/CE/SE

PhD in ML Systems/GPUs

Outils

Nsight Systems/Compute

CUDA

NCCL

Docker

Kubernetes

Slurm

Linux namespaces

cgroups

Triton

MLIR/LLVM

XLA

Description du poste

NVIDIA in France is seeking a senior ML systems engineer to advance large-scale AI inference on GPU hardware. You will contribute features to vLLM, optimize inference pipelines, and help push the pareto frontier of performance.

You’ll design and benchmark GPU kernels, build high-level DSLs, and orchestrate containerized deployments on GPU clusters across clouds. Strong Python/C++ skills and deep knowledge of ML systems are required.