AI Compute Engineer – GPU Infrastructure Architecture

About the role

We are building a next‑generation AI infrastructure platform that delivers high‑performance bare‑metal GPU compute for large‑scale training and inference. As an AI Compute Engineer you will lead the architecture, qualification, and operational readiness of GPU infrastructure across rapidly scaling data centers.

Key responsibilities

• Define architecture standards for NVIDIA HGX platforms, including firmware, BIOS, BMC, OS, drivers, CUDA and NCCL baselines.
• Design and execute qualification and burn‑in strategies for new clusters, GPU platforms and repaired nodes.
• Analyze fleet‑wide performance data, isolate systemic failures, and lead root‑cause analysis for hardware, firmware, software, power and thermal issues.
• Validate real training and inference workloads, set reproducible throughput, scaling and latency baselines, and pinpoint sources of degradation.
• Collaborate with networking, storage, DevOps and facilities teams to route issues with solid evidence and ensure integrated production readiness.

Required profile

• Strong hands‑on experience with NVIDIA GPU infrastructure (HGX or comparable large GPU servers).
• Proven track record designing and leading burn‑in/qualification for multi‑node GPU clusters (H100, H200, B200, B300 or equivalents).
• Deep Linux systems administration and troubleshooting skills.
• Ability to communicate technical findings, risks and remediation plans clearly to engineering, vendors and leadership.

Required skills

• NVIDIA drivers, CUDA, NCCL, DCGM.
• Firmware, BIOS, OS and software baseline definition.
• Root‑cause analysis across GPU, NVLink, NVSwitch, PCIe, network, storage and power domains.
• Python and Bash scripting for troubleshooting and automation.
• Experience with high‑performance fabrics (InfiniBand, RDMA, GPUDirect).