My NVIDIA Platform Engineer interview experience, sharing a short version.
Round 1: Kubernetes/platform fundamentals — architecture, OIDC/RBAC, AWS workload identity, Terraform vs Helm, Argo CD, Controller vs Operator, Vault.
Round 2: Platform design — shared vs dedicated clusters, multi-tenancy, RBAC, quotas, NetworkPolicy, EKS/GKE, upgrades, Terraform state, observability.
Round 3: Troubleshooting — resource requests/limits, QoS classes, pod eviction and enforcing resource policies.
Round 4: GitOps/platform scenarios — multi-cluster management, Argo/Flux, drift, Helm, developer self-service and platform standards.
Round 5: Overall SRE/platform experience, troubleshooting, design decisions and scaling.
Overall difficulty: medium-hard. Most questions had follow-ups like “why this design?”, “what if it fails?” and “how would you scale it?”
Compensation:
https://leetcode.com/discuss/post/8550170/nvidia-vs-atlassian-10-y-experience-help-p0qq/
