Focus on proven cloud-native skills, hands-on troubleshooting, and clear expectations to attract and retain effective platform engineers.

What Matters in Production

Platform engineers thrive when they can diagnose, repair, and prevent infrastructure failures. Hiring for theoretical knowledge or certifications alone fails under pressure. Prioritize candidates who:

• Debug production issues without supervision
• Write maintainable infrastructure-as-code (IaC)
• Understand Observability (metrics, logs, traces)
• Collaborate with developers and SREs

Actionable Hiring Workflow

1. Define requirements with explicit weightings:

   • 40% Kubernetes/OpenShift troubleshooting (e.g., debugging failed deployments, network policies)
   • 30% IaC proficiency (Terraform, Ansible, Operator SDK)
   • 20% Observability tooling (Prometheus, Grafana, OpenTelemetry)
   • 10% Collaboration and documentation

2. Practical assessment:

   • Provide a broken cluster (e.g., misconfigured ingress, failing CSIs)
   • Task: Diagnose root cause, fix, and document steps
   • Evaluate for:
     • Methodical logging/grep usage
     • kubectl describe, logs, exec proficiency
     • Awareness of cascading failures

3. Scenario-based interviews:

   • “How would you recover etcd after a disk failure?”
   • “Explain how you’d automate rollback for a flaky Helm chart”

4. Reference checks:

   • Ask for specific examples of production firefighting

Policy Example: Hiring Rubric

Tradeoff: Strict pass thresholds may reduce candidate pool size but ensure baseline reliability.

Tooling for Evaluation

Use real-world tools to test practical skills:

• Cluster access: Provide a sandbox OpenShift/Kubernetes cluster with intentional failures
• Debugging: Require use of kubectl, curl, grep, jq
• IaC testing: Evaluate Terraform plans or Ansible playbooks for idempotency
• Observability: Task candidates with creating dashboards or alerts from raw metrics

Caveat: Over-reliance on tool-specific questions (e.g., “How do you use Lens?") filters for familiarity, not problem-solving ability.

Troubleshooting Common Hiring Failures

• Symptom: Candidates ace theory but can’t debug a pod in CrashLoopBackoff

  • Fix: Prioritize practical tasks over whiteboard architecture diagrams

• Symptom: Poor team fit due to unclear expectations

  • Fix: Involve SREs and developers in interviews to assess collaboration

• Symptom: High attrition after onboarding

  • Fix: Provide clear career paths (e.g., “Senior Platform Engineer” with incident command responsibilities)

• Symptom: Overemphasis on “10 years of Kubernetes”

  • Fix: Value recent, relevant experience (e.g., multi-cloud deployments, air-gapped clusters)

Final Checklist

• Defined measurable, skill-based criteria
• Tested troubleshooting under time pressure
• Evaluated documentation and communication
• Avoided “culture fit” bias toward diversity of problem-solving approaches

Hiring is a production system—optimize for reducing mean time to recovery, not just resume keywords.

Source thread: Monthly: Who is hiring?