DevOps is a way of working in which the people who build software also ship and run it, with automation in between. DevSecOps adds security to that loop so it happens continuously, not in a review before release. Neither is a single job. The list that recruiters copy into job posts — several languages, several clouds, containers, infrastructure-as-code, CI/CD, GitOps, observability and a long tail of security scanners — describes a platform team. Splitting it by role makes hiring realistic and shows where security responsibilities actually sit.
1 · One Wish-List, Six Roles
| Role | Owns | Typical tools from the list | Security duty |
|---|---|---|---|
| Platform engineer | Kubernetes, container platform, internal developer tooling | Kubernetes, Docker, OpenShift, Helm, Argo CD, Flux | Cluster hardening, admission policies, signed images only |
| Cloud / infrastructure engineer | Accounts, networks, infrastructure-as-code | AWS, Azure, GCP, Terraform, AWS CDK, Bicep, Linux and Windows administration | Least-privilege IAM, IaC scanning, cloud posture |
| CI/CD engineer | Build and release pipelines | Jenkins, GitHub Actions, Azure DevOps | Pipeline secrets, runner isolation, branch protection, artefact signing |
| SRE / observability | Reliability, monitoring, incident response | Prometheus, Grafana, ELK, Datadog, Splunk | Security logging, alert routing to the SOC, runbooks |
| Automation developer | Tooling and glue code | Go, Python, Bash, PowerShell | Secure coding in the tooling itself; no secrets in scripts |
| DevSecOps / product security | Security checks across the pipeline | SonarQube, Trivy and the “twenty other DevSecOps tools” | Choose, tune and own the checks; triage findings with developers |
▸ One strong engineer can be deep in one or two of these roles and conversant in the rest. That is a realistic hire. Asking for depth in all six is how roles stay open for months.
2 · Where Each Security Check Sits
| Stage | Check | Example tools | Owner |
|---|---|---|---|
| Plan | Threat modelling for new features | Whiteboard, OWASP guidance | Product security with the team |
| Code | Secret scanning; static analysis (SAST) | Gitleaks, SonarQube, Semgrep | Developers, tuned by DevSecOps |
| Build | Dependency scanning (SCA) and SBOM | Trivy, OWASP Dependency-Check | CI/CD engineer |
| Build | Container image and IaC scanning | Trivy, Checkov | Platform and cloud engineers |
| Test | Dynamic testing (DAST) against staging | OWASP ZAP, Nuclei | DevSecOps |
| Release | Signed artefacts and provenance | Sigstore / cosign, SLSA practices | CI/CD engineer |
| Deploy & operate | Admission policy, runtime detection, cloud posture | Kyverno or OPA Gatekeeper, Falco, cloud-native posture tools | Platform engineer and SRE |
The free tools in this pipeline are covered in Open-Source Security Stack. Fail builds only on high-confidence, fixable findings; everything else becomes a ticket, or developers learn to ignore the pipeline.
3 · A Realistic Learning Path
- Foundations — Linux, networking, Git, and one scripting language (Python or Bash) properly.
- One cloud, deeply — identity and access, networking and the core services of AWS, Azure or GCP; the others then map across.
- Infrastructure as code — Terraform plus that cloud’s own tool.
- Containers and Kubernetes — build images, run a cluster, understand its security model.
- One CI/CD system end to end, with secrets handled correctly.
- Observability — metrics, logs and traces with one stack.
- Security in the loop — add the checks above to your own project and fix what they find.
4 · Writing a Job Description One Person Can Fill
| Instead of… | Write… |
|---|---|
| A list of forty tools | The two roles from the table you need most, and the outcomes you expect in the first six months |
| “Expert in AWS, Azure and GCP” | “Deep in the cloud we run; able to learn others” |
| “Knows twenty DevSecOps tools” | “Has added security checks to a pipeline and worked findings to closure with developers” |
| “Owns everything in production” | The on-call rotation, team size and what is already automated |
The offensive and defensive security careers that sit next to these roles are in Red, Blue & Purple Teams.
Bottom line
- The wish-list is a team — platform, cloud, CI/CD, SRE, automation and security.
- Hire for depth in one or two roles and breadth across the rest.
- Security belongs at every stage, owned by the people who run that stage.
- Fail builds on what is real and fixable; ticket the rest.
- Learn one cloud and one pipeline deeply before collecting tools.
Sources & method. Tool names from the widely shared “Dear recruiters” post and common job advertisements; capabilities from each project’s documentation. Role split, pipeline placement and guidance are the author’s analysis, consistent with OWASP DevSecOps guidance and SLSA. Tools are examples, not endorsements. Original analysis; no third-party graphics or text reproduced.
← Back to Cyber security