GPU HPC Fleet Reliability Engineer

Coreweave; Bellevue (WA); On-site; USD 83,000 - 110,000; Full time

Benefits offered by this job

Health insurance

Life Insurance

Tuition Reimbursement

Flexible PTO

401(k) with employer match

Catered lunch daily

Job summary

CoreWeave is seeking a Fleet Reliability Operations Engineer to manage and troubleshoot a growing fleet of GPU servers across data centers. You’ll configure, validate, and maintain large HPC clusters, work with cross-functional teams, and drive uptime and efficiency.

This role emphasizes problem solving, documentation, and a proactive approach to incident response in a fast-paced AI infrastructure environment.

Qualifications

• Strong understanding of Linux system administration and internals.
• Ability to troubleshoot hardware and software issues reliably.
• Scripting knowledge in Bash, Python, or PowerShell.

Responsibilities

• Configure and maintain large-scale high-performance supercomputing clusters running state-of-the-art GPUs.
• Troubleshoot hardware and software; coordinate with data center, network, hardware, and platform teams to drive resolution.
• Monitor system performance and take remediation actions for cloud health.
• Collaborate to improve team processes and efficiency.
• Participate in on-call rotations including after hours and weekend work.
• Create and maintain documentation of team processes, knowledge and best practices for system management.

Skills

Linux admin

Troubleshooting

Bash scripting

Python

PowerShell

Education

Bachelor’s degree or equivalent

Tools

Grafana

Prometheus

PromSQL

Kubernetes

Job description

CoreWeave is seeking a Fleet Reliability Operations Engineer to manage and troubleshoot a growing fleet of GPU servers across data centers. You’ll configure, validate, and maintain large HPC clusters, work with cross-functional teams, and drive uptime and efficiency.

This role emphasizes problem solving, documentation, and a proactive approach to incident response in a fast-paced AI infrastructure environment.