GPU Cluster Engineer

Job Summary

To build high-performance GPU clusters, the contract GPU Cluster Infrastructure Engineer will manage design reviews, lead acceptance testing, and establish operational foundations for GPU capacity, working remotely or onsite in New York or San Francisco.

Key responsibilities

• Review cluster designs and bills of materials to identify gaps before hardware ordering
• Lead acceptance testing, including validating cabling, running burn-in, and ensuring vendor deliverables
• Integrate hardware, fabric, and storage telemetry into the observability stack with alerting and automated health checks

Required qualifications

• 3+ years of experience building and operating NVIDIA HGX or DGX clusters in production settings
• Hands-on experience with InfiniBand, including subnet management and fabric bring-up
• Familiarity with GPU node bring-up processes and parallel storage solutions like WEKA or Lustre
• Ability to work effectively both on-site and remotely, including directing remote hands
• Experience with automation tools such as Ansible and monitoring tools like Prometheus/Grafana is a plus