GPU AI Kernel Performance Engineer

AMD; Helsinki; Hybrid; EUR 100,000 - 140,000; Full time

Job summary

AMD is seeking a GPU Performance Engineer focused on AI kernels to optimize key AI workloads on latest GPUs. You will work with a world-class team, shaping high-performance code and hardware-aware software for AI and HPC.

The ideal candidate has 5+ years in GPU software with HIP/CUDA/OpenCL, C++17+, ML frameworks, and strong English communication. You’ll improve kernels, pipelines, and CI, collaborating across hardware and software groups.

Qualifications

• 5+ years of GPU software development experience.
• Strong knowledge of GPU architectures and performance optimization.
• Experience with ML frameworks such as PyTorch and MIOpen is a plus.

Responsibilities

• Continuously optimize AI kernels for AMD GPUs.
• Keep up with GPU architecture and ML algorithm trends.
• Improve development workflows and CI infrastructure for faster deliveries.
• Design and develop GPU and ML technologies.
• Debug issues and implement efficient alternatives.
• Build strong technical relationships with teams and partners.

Skills

HIP/CUDA/OpenCL

C++17+

GPU architectures

PyTorch

MIOpen

Triton

CUTLASS

English communication

Hardware ISA

Education

Bachelor's or Master's in CS/CE/EE

Tools

CI/CD pipelines

Profiling tools

Job description

AMD is seeking a GPU Performance Engineer focused on AI kernels to optimize key AI workloads on latest GPUs. You will work with a world-class team, shaping high-performance code and hardware-aware software for AI and HPC.

The ideal candidate has 5+ years in GPU software with HIP/CUDA/OpenCL, C++17+, ML frameworks, and strong English communication. You’ll improve kernels, pipelines, and CI, collaborating across hardware and software groups.