Hybrid GPU Fleet Operations Engineer

Crusoe

San Francisco (CA)

Hybrid

USD 215,000 - 260,000

Full time

14 days+

Get more replies from employers

Send a job-specific resume in minutes.

Benefits offered by this job

Hybrid work schedule

Industry competitive pay

Restricted Stock Units

Health insurance

401(k) with match

Job summary

Crusoe is seeking a GPU Fleet Operations Engineer to join our Fleet Operations team. You will diagnose, maintain and repair high-density GPU clusters across our data center fleet, ensuring uptime and performance.

The role requires hands-on hardware troubleshooting, firmware updates, and close collaboration with data center operations and engineering. The ideal candidate has strong Linux skills, experience with NVIDIA A100/H200 GPUs, and proficiency in Go, with a track record of working in

Qualifications

Hands-on experience diagnosing and repairing GPU rack-mounted hardware.
Experience with NVIDIA A100, H200, GB200, B200, B300, and AMD 350X/355X.
Strong Linux experience and command-line diagnostics.
Ability to work in fast-paced data center environments.

Responsibilities

Automate deep-level diagnosis and troubleshooting of hardware faults within GPU racks and high-density compute systems.
Develop software to troubleshoot and support GPU platforms including NVIDIA A100, H200, GB200, B200 and AMD 350X / 355X.
Execute component-level diagnosis and remediation for failed or degraded hardware.
Partner with data center operations to manage and perform FRU repairs for GPUs, power supplies, cooling systems, interconnects, and networking hardware.
Conduct post-repair validation, burn-in testing, torch testing, and NVIDIA NCCL testing to ensure system stability and performance.
Implement preventative maintenance procedures to improve fleet reliability and extend hardware lifespan.
Perform firmware and BIOS upgrades across the GPU fleet.
Maintain detailed documentation of maintenance activities, failures, and resolutions in ticketing and asset management systems.
Develop and update standard operating procedures (SOPs) for troubleshooting, repair, and validation workflows.
Collaborate with engineering, software, and data center operations teams to identify root causes of systemic failures and implement preventative solutions.
Participate in a rotating infrastructure on-call schedule (about one week every 4–6 weeks) with daytime coverage and handoff to the Europe team.