GPU / Kernel Performance Engineer (AI)

Location: Mountain View, California, United States; work arrangement: onsite.

Salary: $200,000–$350,000 per year.

Company

Develops intelligent software that analyzes AI workloads and improves their execution on modern accelerator hardware.

What you will do

• Profile GPU workloads to identify bottlenecks, hardware limitations, and efficiency opportunities.
• Optimize existing GPU kernels and develop new kernels for improved workload performance.
• Analyze GPU architecture, memory access, parallel execution, synchronization, occupancy, kernel fusion, and data movement.
• Build reproducible benchmarks, validate numerical correctness, and measure performance consistency.
• Collaborate with ML infrastructure, serving, and applied machine learning teams to translate kernel improvements into system-wide gains.
• Investigate interactions across GPU kernels, compilers, runtimes, and multi-GPU communication and memory bandwidth.

Requirements

• Hands-on experience developing, modifying, or optimizing GPU kernels using CUDA or comparable GPU/HPC programming technologies.
• Strong understanding of GPU architecture, including memory hierarchies, parallel execution, synchronization, and hardware utilization.
• Experience profiling GPU workloads, identifying performance bottlenecks, and measuring improvements with benchmarks or profiling tools.
• Ability to explain the relationship between GPU hardware constraints, software implementation choices, and measured performance.
• Strong programming fundamentals and experience debugging low-level performance issues.
• Willingness to work across GPU kernels, compiler/runtime behavior, and broader software systems.
• Strong technical problem-solving skills and ability to work effectively in a fast-moving engineering environment.

Nice to have

• Experience with CUDA, Triton, CUTLASS, or cuDNN.
• Familiarity with PTX, SASS, NVIDIA Nsight Systems, or Nsight Compute.
• Knowledge of roofline analysis, Tensor Cores, warp specialization, memory coalescing, occupancy optimization, HBM, shared memory, caches, and memory bandwidth limitations.
• Experience with NCCL, NVLink, NVSwitch, AMD ROCm/HIP, TPU/XLA, or performance optimization across multiple accelerator architectures.
• Background optimizing AI training, inference, or high-performance computing workloads.

Culture & Benefits

• Work on core performance optimization technology for next-generation AI infrastructure.
• Direct ownership of GPU kernel and system-level performance improvements.
• Collaboration with experienced engineers from AI hardware and accelerator development teams.
• Small, highly technical engineering team where individual contributions directly affect product performance.
• Exceptional graduates and early-career engineers may take on challenging, high-impact technical work.