A practice-first route into DevOps, whether you're starting from scratch or moving over from development. Every topic gives you one course to learn from, one place to practise, and a portfolio project to prove it.
What you get out of this
Follow this path and you finish with three deployed projects on your GitHub, two or three respected certifications, and — more importantly — the ability to explain how software actually gets from a laptop into production. That combination is what gets juniors interviews.
Give it 10–15 hours a week and the path takes most people 8–12 months. You don't need to wait until the end to start applying: begin around month six, because early interviews show you exactly what to work on next. Plenty of people make this move every year, and almost none of them had a special background — they just kept showing up to the terminal.
Step one
Choose how you'll learn
Pick one main resource and stay with it. Jumping between courses is the most common way people lose months. Whichever you choose, the hands-on labs listed throughout this page are free and worth using alongside it.
All-in-one
KodeKloud
One subscription, a guided DevOps Engineer track, and browser-based labs attached to nearly every lesson — practice is built in rather than something you arrange yourself. Also the standard preparation route for the CKA later in this path.
DevOps Engineer path → Annual plans are usually discounted; they run sales and free-access weeks several times a year, so check before paying full price.
Video-first
TechWorld with Nana
One instructor, one storyline, from Linux through Docker, CI/CD, cloud, Terraform and Kubernetes. Concept-first teaching with unusually clear diagrams. Choose this if you want to understand why before touching a terminal.
DevOps Bootcamp → One-time purchase. Her free YouTube courses are used throughout this page too.
Low or no cost
Free + Udemy mix
Free YouTube full courses and official docs for learning, free hands-on labs for practice, and a few Udemy courses picked up on sale where the free options are weak. This route costs very little and teaches the same skills.
1 · LEARN — the course. Where several are listed they are alternatives: pick the one that matches your budget and learning style, don't do all of them.
2 · PRACTICE — where you actually type commands. Do this while you learn, not afterwards.
3 · REFERENCE — official documentation. Not to be read cover to cover; it's where you look things up, and the habit that makes you employable.
Tags: free costs nothing · udemy cheap when on sale, which is most of the time — never pay list price · subscription included in a paid platform.
The path · check items off as you finish them
Seven stages
Developers switching over can move through Stage 0 quickly — but don't skip Linux and networking. They're where self-taught engineers are most often caught out in interviews.
Stage 0 · Months 1–2pending
Foundations
Everything else runs on Linux, talks over a network, and lives in Git. Rushing this stage is the most common reason people stall later at Kubernetes.
Stage 1 · Months 2–5pending
Containers, pipelines and one cloud
The working core of the job. This stage ends with Portfolio Project #1, which becomes your first real interview story.
Stage 2 · Months 5–9pending
Kubernetes and infrastructure as code
The employability core — the stage that separates people who get interviews from people who don't. Ends with Portfolio Project #2.
Stage 3 · Months 9–12pending
What sets you apart
Most self-taught candidates stop after Kubernetes. These are the topics that make an interviewer lean forward. Ends with Portfolio Project #3, the capstone.
Stage 4 · Every week, from day onepending
The practice layer
Practice is most of this job. These run in parallel with everything above — a weekly routine, not a phase you complete.
Stage 5 · Start at month sixpending
Getting hired
Start applying while you're still learning. Early interviews are free diagnostics — they tell you exactly which gaps to close, months before you'd otherwise find out.
Stage 6 · Read once, then move onpending
What not to spend your time on
Your scarcest resource is hours. These still appear in older tutorials and roadmaps — know they exist, then invest elsewhere.
skipDeep Jenkins and Groovy pipeline authoring
One evening of Jenkins is worthwhile because large organisations still run it. Weeks of Groovy scripting are not — new pipelines are written elsewhere.
skipPuppet and Chef
Legacy maintenance only. If a specific job needs them, you'll pick them up on the job in a fortnight.
skipMemorising command syntax
Assistants and documentation handle syntax. Interviews test whether you understand what a command does and can debug it when it fails at 3am.
skipA second and third cloud before the first is solid
One cloud known deeply beats three known shallowly, and the concepts transfer. Add another after you're hired.
skipOpenStack and on-premise virtualisation
A shrinking niche. Skip unless you're targeting a specific employer that runs it.
skipCollecting more certifications instead of building
Three certifications plus three real projects beats seven certifications and nothing deployed. Interviewers can tell the difference in about four minutes.