Senior GPU Kernel Engineer – Triton, Hybrid

AMD; United States; Hybrid; USD 180,000 - 260,000; Full time

Benefits offered by this job

Benefits at a glance

Hybrid work model

Job summary

AMD is seeking a skilled GPU kernel engineer to design and optimize high-performance Triton kernels for AI workloads. You will collaborate with research, compiler, and hardware teams to push the limits of AMD Instinct GPUs across ROCm and LLVM AMDGPU backends.

The role requires deep expertise in SIMT programming, CUDA-like concepts, and a strong background in GPU architecture. A bachelor’s or master’s degree is expected, and the position is hybrid in the United States.

Qualifications

• Deep experience in GPU kernel development for AI/ML workloads.
• Strong hands-on Triton experience with custom matmul and attention kernels.
• Understanding of modern GPU architectures and memory hierarchy.

Responsibilities

• Design, implement, and optimize Triton/Gluon kernels for large-scale AI workloads.
• Own Triton AMD backend integration across ROCm and LLVM AMDGPU stacks.
• Profile, debug, and drive performance improvements across memory, compute, and scheduling.

Skills

Triton kernel

SIMT programming

GPU architecture

Performance engineering

C/C++

Education

Bachelor's or Master's in CS/CE

Tools

Triton

LLVM

ROCm

MLIR

Job description

AMD is seeking a skilled GPU kernel engineer to design and optimize high-performance Triton kernels for AI workloads. You will collaborate with research, compiler, and hardware teams to push the limits of AMD Instinct GPUs across ROCm and LLVM AMDGPU backends.

The role requires deep expertise in SIMT programming, CUDA-like concepts, and a strong background in GPU architecture. A bachelor’s or master’s degree is expected, and the position is hybrid in the United States.