Data center jobs · AI infrastructure
GPU Cluster Engineer job
What a GPU cluster engineer does day to day, the on-call and hybrid schedule, the Linux, Slurm, Kubernetes and CUDA skills, and how HPC sysadmins, SREs and operations engineers get hired.
By Czarina Tabayoyong, Founder, Critical Path Hiring
Last reviewed · First published
Reviewed as terms, pay and markets change.
The job
A GPU cluster engineer keeps the GPU clusters in an AI data center running. They manage the Linux systems, the cluster scheduler and the NVIDIA drivers, and run hardware diagnostics on the nodes.
A typical day
- Work on Linux systems and cluster schedulers such as Slurm or Kubernetes
- Manage NVIDIA drivers and CUDA
- Run hardware diagnostics on cluster nodes
- Take the on-call rotation
- Visit the site when hardware needs attention
Location and schedule
Often hybrid with site visits, plus an on-call rotation.
Skills and certifications
Required
- Linux
- Slurm or Kubernetes
- NVIDIA drivers and CUDA
- Hardware diagnostics
Routes in
HPC sysadmins, SREs and data center operations engineers move into this role.
Next roles
This role leads to staff engineer or infrastructure manager.
Pay
Pay varies by market and level, and the current ranges are in our data center salary guide.
Questions
- What is the schedule for a GPU cluster engineer?
- An on-call rotation, often hybrid with site visits.
- What does a GPU cluster engineer need to get hired?
- Linux, Slurm or Kubernetes, NVIDIA drivers and CUDA, and hardware diagnostics.
- How do I become a GPU cluster engineer?
- HPC sysadmins, SREs and data center operations engineers move into the role. Staff engineer and infrastructure manager are the next steps.
Next steps
- Join the talent network and tell us your discipline, market and availability
- Look up a term in the data center glossary
- Back to data center jobs explained