Job Description

The Work

AWS customer develops and manages several HPC clusters across AWS, CoreWeave, GCP, and other providers. Several thousand GPUs today scaling to 10X in 2026 and beyond.

This role is Kubernetes-heavy. You'll operate multi-cloud platform infrastructure where misconfigurations or failed upgrades translate directly into thousands of lost GPU-hours. The clusters are large enough that novel failure modes are routine.

• Operate Kubernetes platforms (EKS, CKS, GKE) at significant scale across providers. You're responsible for cluster lifecycle, node pool management, networking policy, and maintaining stability during rapid growth.
• Provision HPC infrastructure through CI/CD system across AWS, CoreWeave, GCP, and OCI, with additional providers to be expanded in the near future.
• Manage job scheduling to allocate GPU compute across training and inference workloads.
• Define and maintain SLIs/SLOs. Build monitoring and alerting. Participate in severity escalation response and author post-incident reviews.
• Coordinate daily with Networking, Storage, Security, and AI/ML platform teams.

• 4+ years in infrastructure engineering, cloud platforms, or HPC.
• Kubernetes is the core requirement. You should have hands-on experience operating clusters at meaningful scale: node pool sizing, scheduler debugging, CNI troubleshooting, and rolling upgrades across large fleets. Candidates whose K8s experience is limited to small or local environments are unlikely to be a fit.
• Terraform proficiency. You'll write and review infrastructure-as-code daily.
• Working knowledge of AWS (EC2, S3, EFS, FSx for Lustre).
• Python for tooling and automation.

Job Tags

Similar Jobs