Platform Engineer Interview Prep Guide

Cover the platform components: service catalog (Backstage), templated service creation with golden paths, automated CI/CD pipelines, environment provisioning with Terraform/Crossplane, observability dashboards auto-configured per service, and a self-service portal. Discuss the principle of golden paths: make the right thing the easy thing. Address how you balance flexibility with standardization and how you handle escape hatches for teams with unique requirements.

Use a modular pipeline architecture: shared pipeline libraries for common steps (build, test, scan, deploy), language-specific builders, and deployment strategies (blue-green, canary) as reusable components. Implement pipeline-as-code with standardized configuration. Discuss caching strategies for build speed, parallel testing, artifact management, and how to handle pipeline failures with automatic rollback. Cover security scanning integration and compliance gates.

Terraform is declarative IaC with HCL, state management, and plan/apply workflow. Crossplane is a Kubernetes-native infrastructure controller using CRDs and the reconciliation loop. Choose Terraform for teams comfortable with its workflow and existing Terraform modules. Choose Crossplane when you want Kubernetes-native infrastructure management, self-service with RBAC via Kubernetes, and continuous reconciliation without state files. Discuss how they can complement each other.

Start with data: survey developers, measure deployment times, track platform adoption metrics, and observe developers using the platform. Identify the highest-friction points. Implement quick wins (better documentation, faster build caches, streamlined onboarding) while working on structural improvements (self-service portal, better error messages, automated troubleshooting). Establish a developer experience metric (DORA metrics, developer satisfaction surveys) and track improvements over time.

Define a Custom Resource Definition (CRD) for your application type. Implement the reconciliation loop that creates/updates Deployments, Services, HPA, and PodDisruptionBudgets based on the custom resource spec. Handle status updates, error conditions, and finalizers for cleanup. Discuss the operator pattern, controller-runtime framework, and how to test operators with envtest. Show awareness of leader election for HA operator deployments.

Use a mono-repo for platform configuration and per-team repos for application manifests. ArgoCD or Flux for continuous reconciliation. Discuss app-of-apps pattern for managing multiple ArgoCD applications, Kustomize or Helm for environment-specific configuration, and how to handle secrets (Sealed Secrets, External Secrets Operator). Cover the PR-based workflow: developers push manifests, automated validation runs, approval gates, and ArgoCD syncs.

Describe the platform with specific capabilities: self-service provisioning, automated deployments, or integrated monitoring. Provide metrics: deployment frequency improvement (daily to hourly), onboarding time reduction (weeks to hours), lead time for changes, and developer satisfaction scores. Explain how you gathered feedback, iterated on the platform, and handled adoption challenges with teams resistant to changing their workflows.

Cover namespace-per-team isolation with network policies, resource quotas, and LimitRanges. Use RBAC for access control, OPA/Gatekeeper for policy enforcement, and Hierarchical Namespaces for team sub-namespaces. Discuss node pools for workload isolation, priority classes for resource preemption, and how to handle noisy neighbors. Address shared platform services: ingress controller, service mesh, monitoring stack, and log aggregation.