There's no degree in DevOps, no standard route in, and no shortage of conflicting advice online. Here's what we actually see working based on the engineers we place and the briefs hiring managers give us.
Ask ten DevOps engineers how they got into the field and you'll get ten different answers. That's not a flaw in the profession, it's the nature of it. DevOps sits at the intersection of software engineering, infrastructure, and operations, and people arrive from all three directions.
What we can tell you, from the roles we fill and the CVs that get interviews, is which routes are working in 2026, which skills employers are genuinely asking for, and where people waste time.
If you haven't already, it's worth reading our companion piece first: What Does a DevOps Engineer Actually Do in 2026? It covers the day-to-day reality of the role. This post covers how to get there.
The three routes we actually see
1. From software engineering. The most common route into the strongest DevOps roles. If you're already a developer, you have the biggest single advantage: you can code properly. Modern DevOps is infrastructure-as-code, automation pipelines, and internal tooling, all of which reward genuine software engineering ability. Developers usually cross over by taking ownership of their team's CI/CD, deployments, or cloud infrastructure, then formalising that into a platform or DevOps role.
2. From sysadmin or IT operations. The traditional route, and still a valid one but the bar has moved. Employers no longer hire "sysadmins who've learned some scripting." They expect ops people to have made a real transition: comfortable in a language like Python or Go, fluent in Terraform or similar, and thinking in terms of automation rather than manual configuration. If you're on this path, the single biggest lever is your coding ability.
3. From graduate or junior positions. The hardest route, because DevOps is fundamentally a role about operating systems in production and juniors haven't done that yet. Genuine entry-level DevOps roles are rare. The more realistic path is starting in software engineering, support engineering, or cloud operations, and moving across within 18–24 months. Some larger organisations run platform engineering graduate schemes, but they're the exception.
The skills that actually appear in briefs
We read hundreds of job specs a year. Strip out the wish-list padding and the same core keeps appearing:
Non-negotiable in almost every brief:
- One major cloud platform, properly. AWS still dominates UK briefs, with Azure strong in enterprise and public sector, and GCP in data-heavy companies. Depth in one beats surface familiarity with all three.
- Infrastructure as code. Terraform is the near-universal standard in the roles we fill. If you learn one tool deeply, make it this one.
- Containers and orchestration. Docker as a given; Kubernetes increasingly expected beyond junior level.
- CI/CD. GitHub Actions and GitLab CI come up most often now, with Jenkins still common in established environments.
- A real programming language. Python most commonly, Go increasingly. Bash alone no longer clears the bar for mid-level roles.
Increasingly requested in 2026:
- Platform engineering thinking: building internal developer platforms and golden paths, not just maintaining pipelines. Many of the best-paid "DevOps" roles are now titled Platform Engineer.
- Security awareness: DevSecOps has moved from buzzword to genuine expectation: secrets management, supply chain security, and shifting security left in the pipeline.
- AI-assisted workflows: Employers don't expect you to be an ML engineer, but they increasingly expect familiarity with AI coding tools and, in some briefs, experience running infrastructure for AI workloads (GPU scheduling, model serving). This is one of the fastest-moving parts of the market.
Rarely the deciding factor, despite what the internet says:
- Certifications on their own. An AWS or Kubernetes certification can help a CV get read, particularly for career-changers, but we've never seen one substitute for demonstrable hands-on work. Treat them as a structured way to learn, not a golden ticket.
What a first DevOps CV needs to show
The CVs that get interviews for entry and mid-level DevOps roles have one thing in common: evidence of doing, not just learning.
That looks like:
- Production ownership, at any scale. "Migrated our team's deployments from manual releases to GitHub Actions" beats a list of ten tools every time. Even in a non-DevOps job title, find and own the automation work.
- A small number of tools, demonstrated deeply. A home lab, a personal project deployed on real cloud infrastructure with Terraform, a documented Kubernetes setup - one genuine project outweighs a long skills list.
- Numbers where you have them. Deployment time cut from hours to minutes, incident recovery improved, cost reduced. DevOps is a measurable discipline; measurable claims stand out.
For the broader principles, our Engineering CV Tips for 2026 covers what employers actually look for in technical CVs.
What you can expect to earn
DevOps and platform roles remain among the better-paid engineering positions in the UK, with London and Cambridge commanding a premium and remote roles compressing regional differences. For current figures across levels, see our UK Software Engineer Salary Guide 2026, which includes DevOps and platform engineering benchmarks.
The honest timeline
From a standing start with no technical background: realistically 2–4 years, going via a software or ops role first. From an existing developer or sysadmin position: 12–18 months of deliberately taking on infrastructure and automation work, then moving into a dedicated role.
The engineers who make the transition fastest aren't the ones who collect the most certifications. They're the ones who find the automation problems in their current job and solve them, then have a story to tell in interviews.
Scam Alert