Senior GPU Performance Engineer - AI Workloads - CA, Santa Clara, United States
Originally posted 28 October 2026 by the employer.
AMD Instinct GPU architectures
About the role
AMD is seeking a Senior GPU Performance Engineer to optimize AI workloads and inform hardware and software co-design for AMD Instinct GPU architectures. In this role, you will work across the hardware and software stack to perform GPU kernel-level performance analysis, isolate microarchitectural bottlenecks, and deliver low-level kernel optimizations for foundation model workloads on large-scale GPU systems.
What you'll do
- Profile and optimize large-scale AI workloads, including transformers, multimodal models, diffusion models, and recommender systems, across multi-node, multi-GPU systems. - Perform deep GPU kernel-level analysis to identify bottlenecks across compute pipelines, memory hierarchy, instruction execution, synchronization, and data movement. - Root-cause performance gaps using GPU performance counters, thread traces, ISA/assembly analysis, and targeted microbenchmarks to deliver low-level kernel optimizations. - Collaborate with kernel, compiler, runtime, framework, and distributed systems teams to improve kernel scheduling, memory utilization, and system-level efficiency. - Develop and maintain benchmarks, microbenchmarks, and traces representative of foundation model and broader AI workloads. - Provide kernel- and microarchitecture-level performance insights to AMD Instinct GPU architecture teams to inform future architecture designs.
What you'll need
- MS or PhD in Computer Science, Computer Engineering, or a related technical field. - Hands-on experience in GPU kernel-level performance analysis and low-level optimization using CUDA, HIP/ROCm, or equivalent technologies. - Understanding of GPU microarchitecture, including execution pipelines, SIMD/SIMT scheduling, registers, memory hierarchy, caches, and interconnects. - Experience with performance counters, thread traces, profilers, and ISA/assembly analysis to resolve GPU kernel performance bottlenecks. - Proficiency in C++ and GPU kernel programming, alongside Python for performance analysis and automation. - Familiarity with AI workloads, frameworks, distributed computing, and communication libraries, such as PyTorch, JAX, Megatron-LM, NCCL/RCCL, and MPI. - Experience with Linux OS, container technologies such as Docker, and Git/GitHub.
Nice to have
- Track record of delivering GPU kernel performance improvements through low-level hardware/software analysis and optimization. - Experience working across kernel, compiler/runtime, framework, distributed systems, and GPU architecture teams to resolve complex performance bottlenecks. - Experience analyzing large-scale AI workloads and translating workload-level performance gaps into actionable kernel, software-stack, or microarchitecture improvements.
Skills: GPU kernel, CUDA, ROCm, microarchitecture, profiling, C++
This role has been open 0 days — well below the 55-day median for Infrastructure/Platform Software roles.
Infrastructure/Platform Software · GPU Software Stack
|
Open roles in category
258
|
Median days open
55 d
|
Median salary
$220k
|
See the full market breakdown ▾Category comparison, skills in demand, and who else is hiring
| Metric | AMD | All employers we track in this specialty (258 roles · 19 employers) |
|---|---|---|
| Open roles in this specialty | 55 | 258 |
| Open roles in the wider Software, Firmware & Systems family | 389 | 6395 · 147 employers |
| Median days open | 57 d | 55 d (+2 d vs this employer) |
| Median salary (USD postings) | — | $220k |
Skills observed across this category: GPU kernel, CUDA, ROCm, microarchitecture, profiling, C++
Who's hiring in this category
- NVIDIA · 127 open roles · median 58 d
- AMD (this employer) · 55 open roles · median 57 d
- Intel Corporation · 29 open roles · median 15 d
- Qualcomm · 17 open roles · median 188 d
- Arm Holdings · 8 open roles · median 1 d
- Graphcore · 4 open roles · median 8 d
How we counted: 258 open Infrastructure/Platform Software (GPU Software Stack) roles from 19 employers tracked in the SemiconductorJobs index, counted 29 Oct 2026. Specialty figures count only roles carrying this exact specialty label, so an employer's related work in neighbouring specialties is not included there — it is counted in the wider Software, Firmware & Systems family row. Figures refresh nightly.