Lead HPC & AI Scientist — GPU Optimization & Porting

AMD; Paris; Sur place; EUR 90 000 - 130 000; Plein temps

Avantages offerts par ce poste

AMD benefits

Résumé du poste

AMD's Software and Solutions Team is seeking a Lead HPC and AI Computational Scientist/Engineer to optimize scientific computing applications for AMD GPU platforms. Based in Paris, you will collaborate with academia, industry, and research labs to port, tune, and scale large-scale HPC workloads and AI workflows.

The role requires expert HPC performance engineering, GPU programming (CUDA/HIP/OpenMP), and cross-functional teamwork across distributed teams; on-site presence is required for

Qualifications

• PhD or equivalent experience in HPC or AI GPU domains.
• Minimum 7 years of relevant HPC/AI GPU experience.
• Proficient in French and English communications.

Responsabilités

• Develop, port, and optimize high-performance computing software and applications for AMD hardware.
• Collaborate with COE team to solve issues and share best practices.
• Represent AMD to customers and third parties, acting as customer advocate during presentations.
• Engage with AMD product groups to drive resolution of customer issues.
• Develop and present training materials to internal and external audiences.

Connaissances

HPC performance optimization

GPU programming

Scientific computing

Parallel programming (MPI/OpenMP)

French & English communication

Formation

PhD in Computer Science, Computational Physics, Engineering or related subjects

Outils

CUDA

HIP

OpenMP

Description du poste

AMD's Software and Solutions Team is seeking a Lead HPC and AI Computational Scientist/Engineer to optimize scientific computing applications for AMD GPU platforms. Based in Paris, you will collaborate with academia, industry, and research labs to port, tune, and scale large-scale HPC workloads and AI workflows.

The role requires expert HPC performance engineering, GPU programming (CUDA/HIP/OpenMP), and cross-functional teamwork across distributed teams; on-site presence is required for