GPU Math Libraries Engineer - AI & HPC Emulation

NVIDIA; France; Hybrid; EUR 90,000 - 130,000; Full time

Job summary

NVIDIA is seeking software engineers to advance its math libraries teams for AI and HPC kernel generation, focusing on emulation of math operations across precisions. You will design and optimize dense linear algebra software on GPUs, mentor engineers, and partner with product teams to define roadmaps.

Ideal candidates hold a PhD or MSc in a related field, 5+ years in CUDA/C++ production software, and strong knowledge of GPU architecture. Hybrid work model and collaborative environment offered.

Qualifications

• PhD or Master’s degree in Computer Science, Applied Math, or related field (or equivalent experience).
• 5+ years of experience designing, developing, testing, maintaining, and optimizing production software using CUDA and C++.
• Strong knowledge of GPU hardware architecture (preferable) or CPU.
• Strong fundamentals in finite precision arithmetic and numerical linear algebra.
• Great teamwork, communication, and documentation habits.

Responsibilities

• Scope, design, and implement high-quality numerical dense linear algebra software on GPUs.
• Provide technical leadership and mentorship to library engineers and interns.
• Collaborate with product management and customers to define feature roadmaps and performance goals.
• Identify opportunities to improve performance and reduce maintenance via re-architecting.

Skills

CUDA

C++

GPU architecture

Numerical methods

Team collaboration

Documentation

Education

PhD or Master’s in CS/Math

Tools

CUTLASS

Python

Job description

NVIDIA is seeking software engineers to advance its math libraries teams for AI and HPC kernel generation, focusing on emulation of math operations across precisions. You will design and optimize dense linear algebra software on GPUs, mentor engineers, and partner with product teams to define roadmaps.

Ideal candidates hold a PhD or MSc in a related field, 5+ years in CUDA/C++ production software, and strong knowledge of GPU architecture. Hybrid work model and collaborative environment offered.