Senior GPU AI Software Engineer — Kernels to Scale AI

Advanced Micro Devices, Inc.; Santa Clara (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

Advanced Micro Devices, Inc. is seeking a senior software engineer who blends deep systems performance work with modern AI, spanning GPU kernels through distributed training and inference on ROCm.

You will own the AI software stack, optimize large-scale kernels, collaborate on hardware-software co-design, and help push AMD to the forefront of transformer and LLM workloads. This hands-on, high-impact role also involves mentoring teammates and shaping the roadmap for future accelerators.

Qualifications

• Expert-level C++ in large, performance-critical systems.
• Strong grasp of GPU architecture, memory hierarchies, and kernel optimization.
• Hands-on with C++/HIP/CUDA codebases (ROCm, cuBLAS, cuDNN, NCCL).

Responsibilities

• Own the AI software stack from low-level kernels to large-scale distributed systems.
• Accelerate training, post-training, and inference for LLMs and AI workloads.
• Co-design hardware and software with cross-functional teams and contribute to ROCm ecosystem.

Skills

C++ performance engineering

GPU architecture

HIP/CUDA

Profiling & debugging

Education

Bachelor's in CS/CE or equivalent

Master's preferred; PhD a plus

Tools

ROCm

CUDA ecosystem

Nsight

Job description

Advanced Micro Devices, Inc. is seeking a senior software engineer who blends deep systems performance work with modern AI, spanning GPU kernels through distributed training and inference on ROCm.

You will own the AI software stack, optimize large-scale kernels, collaborate on hardware-software co-design, and help push AMD to the forefront of transformer and LLM workloads. This hands-on, high-impact role also involves mentoring teammates and shaping the roadmap for future accelerators.