Senior Systems Performance Engineer
Core
Develop and tune synthetic GPU workloads and AI training/inference frameworks to stress hardware subsystems, analyze performance anomalies, and build automation for validation of GPU accelerated computing products.
Role type
Senior Systems Performance Engineer (Validation)
Builds
Automated test plans, performance models, and validation coverage for DGX Server products
Domain
Accelerated computing, GPU hardware validation, AI infrastructure
Deliverable
production ML models
Required skills
GPU architecture analysis, performance modeling, CUDA programming, Python, Linux debugging, telemetry analysis, experiment design
Preferred skills
NVIDIA Dynamo, TensorRT-LLM, vLLM, SGLang, CUDA kernel optimization, x86/Arm server familiarity, NVLink/PCIe/InfiniBand knowledge
Technologies
CUDA, PyTorch, TensorRT-LLM, vLLM, SGLang, Linux, Python, NCCL, cuBLAS, CUTLASS
Responsibilities
Develop synthetic GPU workloads to stress hardware subsystems; onboard and adapt AI training/inference workloads to new GPU platforms; apply architectural analysis to develop system-level performance models; design controlled experiments and correlate predictions with telemetry; analyze performance anomalies and hardware/software failures; build automation for workload deployment and analysis
Seniority
Senior, hands-on IC
