Can a self-taught person become a cloud engineer? Yes — with one honest caveat.
The learning is the part self-taught people are best at. It is the two feet after the learning, the portfolio and the job hunt, where most of them quietly stall.
New to cloud? CAMPUX is a free, build-first course. Start here →
Yes — cloud engineering does not require a degree; hiring is by proof, not paper. The part the other guides skip is the harder truth underneath it: certified is not the same as hired. Self-taught people rarely fail at the learning. They stall one step later, at the portfolio and the job hunt, where a stack of certificates meets a hiring manager who wants to see work.
Walk into almost any cloud team and count the degrees in the room. You will find computer science graduates, sure, but you will also find a former help-desk tech, someone who came up through networking, a person who taught themselves during a slow year and never looked back. The field is young enough that the people running it mostly did not train for it in school — there was no school for it when they started. That history is why the door is genuinely open. It is also why the door is crowded, and why "I passed the certs" is where a lot of self-taught candidates get stuck rather than hired.
What actually replaces a degree signal
A degree was never the thing employers wanted. It was a proxy — a cheap way to bet that someone can take on hard, abstract material, stick with it for years, and finish. Most cloud job postings still say "degree or equivalent experience," and that second clause exists so a hiring manager can ignore the first when a candidate shows the real thing directly. Your job as a self-taught candidate is to supply that real thing so plainly that nobody reaches for the proxy.
Three artifacts do the replacing, and they work as a set, not a menu. A role-based certification replaces the "can learn hard things" signal — it tells a recruiter you sat a real exam that assumes hands-on skill and passed. A portfolio of real projects replaces the "can do the work" signal that no exam can carry. And a shipping track record, even a modest one (a repo with commit history, a project you improved over three weekends), replaces the "will finish, not flake" signal. A degree bundled those three into one certificate. You are unbundling them and proving each on its own.
The good news is that this is more honest currency than a diploma. A hiring manager can click your GitHub and see the work; they cannot click your degree and see whether you learned anything. The self-taught path, done right, produces better evidence than the traditional one — which is exactly why it works.
The degree was a bet that you can finish hard things. A portfolio is not a bet. It is the finished thing.
The failure mode nobody addresses
This is the email that arrives, in some form, more than any other: "I did the courses, I passed AZ-900 and AZ-104, I've been applying for two months, and I have not had a single interview. What am I doing wrong?" It is written by someone who did everything the beginner guides told them to do. The guides said get certified and apply. They got certified. They applied. Nothing.
What is going wrong is almost never the learning, and that is the cruel part — the learning was the hard, lonely, months-long grind, and it worked. The stall is one step further along, at the point where studying is supposed to convert into interviews. It does not convert on its own. A certificate proves you learned material; it does not prove you can build and troubleshoot, which is the entire job. When thousands of applicants hold the identical badge, the badge stops being a differentiator and becomes a filter you share with everyone else in the pile. I have written the AZ-900 version of this in detail in passed AZ-900, still not getting interviews, and the diagnosis generalizes: the resume is built around what you studied instead of what you can produce.
Naming the failure mode precisely matters, because it changes what you do next. If you believe the problem is that you have not learned enough, you go buy another course and delay the job hunt by three months. If you understand the problem is that you have not shown enough, you stop taking courses and start building and publishing. The second person gets hired. The first person collects certificates.
Studying feels like progress because it is measurable and safe — a lesson finished, a score improved, a cert earned. Building in public feels like exposure, because the code is imperfect and someone might see it. So the anxious self-taught learner keeps studying long past the point of diminishing returns and never ships. If you have three certs and zero published projects, more study is not your gap. Shipping is. The discomfort is the signal you are finally doing the part that counts.
What a hiring manager screens for beyond the cert
Sit on the other side of the desk for a minute. A manager filling a junior seat has a stack of resumes that all list the same two or three Azure certs. The cert has already done its only job — it got the resume past the keyword filter. Now the manager is looking for the thing that separates one identical-looking candidate from the next, and it is not a fourth certificate.
They are screening for evidence you have touched the real thing. Have you defined infrastructure as code, or only clicked through the portal in a tutorial? Have you ever had a deployment fail and had to read the error, form a theory, and fix it — or has everything you have done worked the first time because a video told you which buttons to press? Can you explain a technical decision you made and why you made it that way rather than another way? Those three, infrastructure you built as code, a problem you diagnosed, a decision you can defend, are what a manager is reading for. None of them appear on a certificate.
This is the artifact set worth building on purpose. A couple of real lab repos with the IaC and a small CI/CD pipeline that deploys them. One honest troubleshooting write-up: here is what broke, here is how I found it, here is the fix — the messier and more real, the better. And a rehearsed ninety-second script explaining one technical decision from your projects, because interviewers will ask, and "I followed the tutorial" is a losing answer. The Azure portfolio guide lays out exactly what to build and how to write it up.
The three self-taught paths that work
"Self-taught" is not one path; it is three, and they suit different starting points. Being honest about which one you are on saves months.
Cert-first. You have no IT job and you are studying from zero. You lead with a role-based cert to get past filters, then spend at least as long building a portfolio as you spent studying. This is the purest self-taught route and the slowest, because you are manufacturing the "experience" signal entirely from personal projects. It works, but only if you are disciplined about shipping and not just accumulating certs.
IT-support entry. You take a help-desk or junior sysadmin role first — sometimes an easier door to open — and move to cloud from inside. This is often the fastest route to a paycheck in the industry, because the support job gives you real production exposure, a professional reference, and the informal credential of "has held a technical job." You study cloud on the side and transition, either internally or to a new employer, once your portfolio is real.
Bootcamp-assisted. Still self-directed, but with a structured curriculum, a project sequence, and a defined endpoint so you do not spend six months deciding what to learn next. The honest trade-off is that a bootcamp does not confer experience or a guaranteed job; what it buys is structure and a portfolio-by-design, which is worth a lot if your failure mode is drifting between free tutorials without ever finishing a real build.
| Path | Typical cost | Typical time to job-ready | What it proves to a hiring manager | Main risk |
|---|---|---|---|---|
| Self-taught | Low — cert exam fees plus a little cloud spend; roughly $150–$500 | ~6–12 months part-time | Self-direction and initiative — you built the whole thing without being told to | No external structure; easy to over-study and never ship a portfolio |
| Bootcamp | Mid — free/low-cost options up to several thousand dollars | ~3–9 months | A structured, finished project sequence and a defined skill baseline | Uneven quality; no bootcamp confers real experience or a guaranteed job |
| Degree (CS/IT) | High — often tens of thousands over years | ~2–4 years | Endurance and fundamentals, but rarely current cloud skills on its own | Slow and expensive; graduates still need certs and a portfolio to land cloud roles |
Read the figures above as direction, not gospel. Costs and timelines swing hard with your starting point, your hours per week, and the market you are applying into. What holds firm is only the ordering: self-taught cheapest and most self-directed, degree slowest and most expensive. What every row has in common is the bottom-right reality: none of these three, on its own, hands you the portfolio. You still have to build that. For a grounded look at the timeline question specifically, how long it takes to get a cloud job without a degree works the numbers harder.
A self-taught Azure roadmap
Azure is a practical target for the self-taught, for one unglamorous reason: its certification ladder maps cleanly onto job titles employers post. "Azure Administrator" is a real, searchable job title, and AZ-104 is the exam named for it. That alignment means a self-taught candidate is never guessing whether their credential matches a role; the role and the cert share a name. The sequence that works looks like this.
Weeks 1–6, foundations and vocabulary. Work through cloud concepts and core Azure services and sit AZ-900. It is a fundamentals exam, proof you can describe the cloud, not build it, so treat it as the on-ramp, not the destination. Its real value is that it lets you learn the rest without tripping over unfamiliar words.
Weeks 6–20, the role-based cert with your hands on the keys. Study for AZ-104, the Azure Administrator Associate, and, above all, do every topic for real in a free Azure account rather than only reading about it. Virtual networks, storage, identity, compute, governance: stand each one up, break it, fix it. You are studying for the exam and building portfolio muscle in the same hours.
Weeks 16–28, build and publish in the open. Turn that hands-on work into two or three finished projects, defined as code and pushed to GitHub with a README each. Add the troubleshooting write-up and the decision script from earlier. This is the phase self-taught learners skip and the phase that gets them hired, so protect it — it is not optional polish, it is the whole product.
Weeks 24 onward, run a real job search. Rewrite the resume around what you built, target "Azure Administrator" and "junior cloud engineer" roles that match your level, and apply deliberately rather than firing into the void. If you want the fuller version of the from-scratch route, how to become a cloud engineer with no experience covers the same ground with the beginner's questions answered.
Questions people also ask
Can you become a cloud engineer without a degree?
Yes. Cloud engineering is one of the more credential-flexible technical fields, and most job postings ask for a degree "or equivalent experience" — a line that exists precisely so a hiring manager can wave it away when a candidate shows real work. The degree was never the point; it was a proxy for "can learn hard things and finish them." A role-based certification plus a portfolio of things you built proves the same thing more directly.
How long does it take to become a self-taught cloud engineer?
For someone studying part-time around a job, roughly six to twelve months to job-ready is a realistic band — a few months to foundations and a role-based cert, then a few more to build and document real projects. People already in IT support move faster because half the groundwork is done. Treat those figures with suspicion as promises and as direction; the variable that moves the timeline is hours per week on hands-on work, not the calendar.
Do certifications get you a cloud job?
They get you past filters, not across the finish line. A certification proves you learned the material; it does not prove you can build and troubleshoot, which is what the job is. Cert-only resumes stall because thousands of people hold the same badge and none of it shows work. The candidates who get hired pair a role-based cert with a portfolio of real projects and can talk through a technical decision they made.
What does a self-taught cloud engineer need in a portfolio?
Evidence you can build and fix, not just describe. Aim for two or three real projects in a live cloud account: infrastructure defined as code (Bicep or Terraform), a small CI/CD pipeline that deploys it, and at least one troubleshooting write-up that walks through something that broke and how you diagnosed it. Publish the code on GitHub with a README per project explaining the goal, the architecture, and what went wrong. The write-up is what separates you from the identical cert pile.
Can a self-taught person get an Azure job?
Yes, and Azure is a practical target for self-taught learners because its certification ladder maps cleanly to job titles employers post. Start with AZ-900 for the vocabulary, move to AZ-104 (Azure Administrator) which assumes hands-on skill, and build your portfolio in a free Azure account. "Azure Administrator" and "junior cloud engineer" are real, postable titles, so a self-taught candidate with the AZ-104 and a portfolio is applying to roles that exist rather than reaching for a senior seat.