DevOps Engineer Hiring Guide
DevOps in 2026 means "owns deployment, infrastructure, observability, and on-call across the SDLC." It is one of the highest-demand, lowest-supply roles in Indian IT staffing. Most "DevOps" CVs are actually system administrators with a Kubernetes certificate. The rubric below separates real platform engineers from CI/CD button-pushers.
Key skills
Must-have
Cloud platform depth
3+ years hands-on with AWS, GCP, or Azure including at least 5 services beyond the basics (e.g., IAM, networking, secrets management, monitoring).
IaC fluency
Terraform, Pulumi, or CloudFormation — has actually written and maintained infrastructure as code, not just edited someone else's modules.
CI/CD pipeline ownership
Has built and debugged pipelines (GitHub Actions, GitLab CI, CircleCI, Jenkins). Understands why fast pipelines matter.
On-call experience
Has been on a real on-call rotation, paged at 3am, and resolved an incident. Without this, "DevOps" is theoretical.
Nice-to-have
Kubernetes operational depth
Has run K8s in production — debugged pod crashes, networking issues, autoscaling. Many "K8s engineers" only deploy YAMLs.
Observability tooling
Datadog, Prometheus/Grafana, Honeycomb, or similar. Has actually configured alerts and built useful dashboards.
Security/compliance awareness
IAM least-privilege, secret rotation, basic SOC2/ISO 27001 controls. Reduces ramp on regulated clients.
Scripting depth
Bash + at least one of Python/Go for tooling. Pure GUI-driven DevOps is a yellow flag.
Interview questions (7)
Walk me through the worst production incident you have been on-call for. What was the trigger, the response, and the postmortem?
What to listen for
Specific timeline. Real diagnosis (logs, metrics, system state). Postmortem with action items. "We never have incidents" is disqualifying for a senior role.
How do you decide when to use Kubernetes vs ECS vs plain EC2?
What to listen for
Pragmatic — based on team size, ops capacity, deployment frequency, blast radius. "Always K8s" or "Never K8s" is dogmatic.
Describe your CI/CD pipeline. How fast is it from commit to production, and where does the time go?
What to listen for
Specific numbers. Awareness of bottlenecks. If the answer is "we deploy weekly with manual approval," they have not optimized for velocity.
How do you handle secrets in your pipelines and runtime?
What to listen for
AWS Secrets Manager / Vault / cloud-native secret store. Not "we use environment variables in the .env file."
What is your approach to alerting? How do you avoid alert fatigue?
What to listen for
SLO-based alerting. Symptom-based, not cause-based. Tiered (page vs ticket vs FYI). Not "we alert on every error."
A developer says their service is slow. Walk me through how you investigate.
What to listen for
Metrics first, then logs, then traces, then code. Not "I would log into the box and run top." (Some senior DevOps will and that is fine — but the modern answer leans on observability.)
How do you balance "move fast" with reliability?
What to listen for
Error budgets, canary deploys, feature flags, gradual rollouts. Real frameworks, not just "we are careful."
Evaluation rubric
Score each candidate against these weighted criteria. Total: 100%.
| Criterion | Weight | Signal |
|---|---|---|
| Cloud + IaC depth | 30% | Multi-year hands-on with major cloud + production Terraform/Pulumi/CloudFormation. |
| On-call + incident response | 25% | Has owned real incidents and authored real postmortems. |
| Pipeline ownership | 20% | Has built and optimized CI/CD pipelines. Knows their commit-to-prod time. |
| Observability discipline | 15% | Has set up alerts that paged them. Has tuned alerts after fatigue. |
| Security mindset | 10% | IAM least-privilege, secret hygiene, awareness of common cloud misconfigurations. |
Red flags
CV is a list of certifications with no project depth
Has never been on a real on-call rotation
Claims Kubernetes expertise but cannot debug a CrashLoopBackOff in 5 minutes
Always-on-call equals 24/7 with no rotation — burnout candidate, will leave fast
Cannot name a specific incident or production debugging story
Apply this rubric automatically with CVPRO
Upload DevOps Engineer CVs and let AI score every candidate against the same 42-point evidence rubric.
Try CVPRO Free