System Software Engineer – Data Center GPU Compute Diagnostics
$152,000–$241,500 year
On-site · Durham, North Carolina, United States
Job Summary
Develop CUDA/C++ diagnostic workloads and GPU compute tests to stress Tensor Cores, SMs, L2/cache hierarchy, HBM memory, and related power/thermal operating points; implement and tune GEMM-style diagnostic workloads with NVLink/PCIe/CPU subsystems; contribute to higher-level AI workload tests using PyTorch-based models to stress GPUs and system software; collaborate with hardware architecture, driver, manufacturing, and field teams throughout the product lifecycle; bring up and validate new hardware features with pre-beta GPU drivers, low-level diagnostic software, and system telemetry; triage and debug ECC, HBM behavior, thermal limits, voltage/frequency margining, and PCIe/NVLink errors.
Required Qualifications
- BS or MS degree in Electrical Engineering, Computer Engineering, Computer Science, or equivalent experience
- 5+ years of system software, GPU software, embedded software, or hardware validation experience
- Experience writing low-level diagnostics, interacting with device firmware and hardware level debuggers
- Strong C/C++ and Python programming skills
- Exposure to GPU architecture, CUDA kernels, GPU compute workloads, or related accelerator programming is strongly preferred
- Working knowledge of memory systems, ECC behavior and DMA engines
- Familiarity with GEMM-style workloads
- Awareness of voltage/frequency characterization, thermal testing, power stress, or related silicon validation concepts such as Vmin/Fmax and P-state testing
- Experience using modern AI development and analysis tools to improve engineering velocity, including code development, debugging, and test creation
- Strong problem solving and low-level debugging skills
- Your base salary will be determined based on your location, experience, and the pay of employees in similar positions. The base salary range is 152,000 USD - 241,500 USD.
Apply with one swipe on Sorce. We auto-fill applications and apply on your behalf — no cover letters, no 40-minute forms.
Hiring someone like this?
Get your role in front of qualified candidates on Sorce.