GPU AI Performance & Benchmark Engineer

Yoh, A Day & Zimmermann Company; California (MO); On-site; USD 250,000 - 300,000; Full time

Benefits offered by this job

Medical+Vision

HSA

Life & Disability

401K

EAP

Direct Deposit

Referral bonus

Training opportunities

Job summary

Yoh, a Day & Zimmermann company, seeks a hands-on Performance/Benchmark Engineer to characterize AI workloads on large-scale NVIDIA GPU systems. The role sits within the NVIDIA GPU Systems architecture team and focuses on GPU compute performance, AI inference, and system benchmarking.

Responsibilities include developing benchmarks, analyzing latency, throughput, and memory behavior, and collaborating across architecture, compute, networking, and software teams to optimize end-to-end performance

Qualifications

• Deep hands-on experience with NVIDIA GPU compute platforms and AI/ML performance benchmarking.
• Strong understanding of AI inference, model performance, and GPU architecture.
• Experience with DGX, B200/B300, H100/H200, Blackwell, Hopper or comparable GPUs.
• Experience analyzing latency, throughput, GPU utilization, memory bandwidth, and scaling.
• Proficient in Python scripting and automation.

Responsibilities

• Develop and execute benchmarks for AI inference on NVIDIA GPU systems.
• Characterize throughput, latency, and memory on DGX and B200/B300 platforms.
• Identify bottlenecks and propose hardware/software improvements.
• Create benchmarking methods, automation, and reporting for repeatable results.
• Collaborate to optimize end-to-end AI cluster performance.

Skills

GPU compute platforms

AI/ML benchmarking

Python scripting

Multi-GPU scaling

System-level benchmarking

Tools

CUDA

Nsight

TensorRT

PyTorch

NCCL

Triton Inference Server

Job description

Yoh, a Day & Zimmermann company, seeks a hands-on Performance/Benchmark Engineer to characterize AI workloads on large-scale NVIDIA GPU systems. The role sits within the NVIDIA GPU Systems architecture team and focuses on GPU compute performance, AI inference, and system benchmarking.

Responsibilities include developing benchmarks, analyzing latency, throughput, and memory behavior, and collaborating across architecture, compute, networking, and software teams to optimize end-to-end performance