Core responsibilities
- Build isolated, reproducible environments
- Manage quotas, GPUs and spending
- Provide self-service deployment paths
Technical skills
- Kubernetes orchestration and GPU scheduling
- IAM, networking and policies
- SLOs, capacity and developer experience
Behavioural skills
- Gather requirements from several teams
- Handle exceptions without weakening the platform
Typical deliverable
Isolated workspaces with accelerator quotas, cost visibility and a shared registry.
Possible assessment methods
Resolve GPU contention on a multi-team platform with clear fairness rules and metrics.
