What skills do you need to become a cloud engineer?
Seven technical skills and one nobody puts on a job post. Certs test whether you can recall them one at a time. Jobs test whether you've ever wired them together into something that deploys and survives.
New to cloud? CAMPUX is a free, build-first course. Start here →
To become a cloud engineer you need seven technical skills — networking and identity, one cloud platform in depth, Linux and scripting, infrastructure-as-code, CI/CD, monitoring, and cost management — plus one soft skill nobody lists: the ability to own an incident. The honest catch is that a certification proves you can recall these separately; a job requires you to have wired them together at least once.
That is the whole answer, and the rest of this note is me defending it. I have hired for these roles and I have sat on the other side of the table trying to get hired for them. The list of skills is not controversial. What trips people up is the shape of the list — which skills carry the interview, which ones are table stakes, and the difference between studying a skill and being able to use it when a director is standing behind your chair asking why the site is down.
The seven, in the order that pays
Not every skill on this list weighs the same. I have ordered them the way they tend to show up in postings and screens, because that is the order you should learn them in if you want to be employable at each step rather than at the end. If you want the demand data broken out by role — cloud admin, cloud engineer, DevOps — our skills radar ranks each skill by how often it appears in current job listings. Here is the stack, and what "job-ready" looks like for each rung.
| Skill area | What "job-ready" looks like | Where CAMPUX trains it |
|---|---|---|
| Networking & identity | You can draw a private network from memory — subnets, a firewall rule, why a resource can or can't reach another — and you understand who is allowed to do what and how that's granted. This is the half most beginners skip and most interviews open with. | Entra & subscriptions, RBAC & Policy |
| One cloud platform, deep | You know one provider's core services — compute, storage, networking, identity — well enough to pick the right one and defend the choice. Depth in one beats a shallow tour of three. We teach Azure; the shape transfers. | What is cloud, Compute, Storage |
| Linux & scripting | You live in a terminal without flinching and can write a script that calls an API, loops over resources, and fails cleanly. Not software engineering — automation glue. Python or Bash, pick one and get fluent. | Scripting for cloud engineers |
| Infrastructure-as-code | You define infrastructure in a file, commit it, and stand it up from nothing with one command. If you're still clicking through a portal to build things, you're doing a task a machine should do. This is the skill that most separates hires from candidates. | Bicep, Terraform |
| CI/CD | You can wire a pipeline that takes a commit and turns it into deployed infrastructure or a running app, with the credentials handled safely. The thing that makes IaC a practice instead of a party trick. | REST APIs & ARM, plus the pipeline classes |
| Monitoring & observability | You can answer "is it healthy, and how would I know before a user tells me?" You set up the alert, read the logs, and find the failing component under time pressure. | Incident response and the monitoring classes |
| Cost & FinOps | You know what your architecture costs before the bill arrives, and you can cut it without breaking it. The skill that gets a junior noticed, because it's the one management feels directly. | Cost management & FinOps |
Notice what's not a separate row: certifications, Kubernetes, and "cloud security" as a standalone. Certs are a way to prove some of these, not a skill of their own. Kubernetes is a deep specialty you reach for after the fundamentals, not a starting skill. And security is not a rung — it runs through every rung, from the identity you configure to the firewall rule you write to the secret you keep out of your pipeline. Treat it as a bolt-on and it will be the thing that fails your interview.
The skill nobody puts in the posting
Here is the one that does not fit in a table because you cannot study it from a book. You can own an incident. Something is broken in production, the dashboards are red, and you are the person who reads the failing system, decides whether to roll back or push through, communicates what you know to people who are scared, and answers for the call afterward.
No job description lists "stays calm when production is down and takes responsibility for the fix." They list the tools. But every senior person interviewing you is quietly testing for exactly this, because the tools are learnable in months and this is the thing that separates someone they can leave alone at 3 a.m. from someone they cannot. The way you build it is by owning small things end to end — a deploy, a broken pipeline, a cost spike you caught — long before anyone hands you production.
The tools are table stakes. Owning the outcome is the job.
Certified tests recall. Hired tests wiring.
This is the gap the competing "top 10 skills" listicles structurally cannot own, because most of them are selling you a cert. So let me say it plainly. A certification exam — AZ-104, AZ-900, the AWS equivalents — tests whether you can recognize the right answer about each of these skills in isolation. It is multiple choice against a clock. It is recall. It is genuinely useful, and it is not the job.
A job screen tests something the exam cannot: whether you have ever taken networking, identity, IaC, and a pipeline and connected them into one working thing that survived contact with reality. That is why the candidate with a wall of badges and no projects loses to the candidate with one messy repository that actually deploys. The badges prove you studied the parts. The repository proves you assembled them, which is the only thing the parts were ever for.
You can pass AZ-104 in a few weeks of focused study — recall is fast to build. Becoming job-ready across these seven skills takes most people six to twelve months of consistent work, faster if you already sit in IT support or a sysadmin seat (as of 2026; the range is what I see from people making the jump, not a survey figure). The exam is the short, cheap part. The wiring is the long part, and it is the part that gets you paid.
Azure-flavored, but the fundamentals transfer
We teach this stack on Azure, and I want to be honest about why and about the limits of that choice. Azure is the safe default for a big share of the corporate job market — Microsoft-heavy enterprises, government, healthcare, finance. If that is where you are aiming, learning Azure is learning the tool your future employer already runs.
But the skills underneath are not Azure. A subnet is a subnet. Role-based access control, a deployment pipeline, a monitoring alert, a cost alert — the concepts are identical across Azure, AWS, and Google Cloud, and only the service names and the exact syntax change. Learn one platform to real depth and the second one is a few weeks of translation, not a fresh start. Anyone who tells you to learn all three at once is wasting your first year. Pick the one nearest the jobs you want, go deep, and let the fundamentals do the porting. If you're still deciding where you fit, the roles page breaks down cloud admin, cloud engineer, and DevOps so you can aim the stack at a specific title.
How to build the list for real
Do not try to learn these in seven separate courses and hope they fuse on their own. They don't. The move is to learn each skill in the context of building one thing, so that by the end you have both the skills and the wiring the job screens for. That is the entire design of this bootcamp — start at what cloud even is, and come out the other side having stood up real infrastructure as code, deployed by a pipeline, monitored, and cost-checked. If you want the skills forced into practice on a clock, the 100 Days of Azure track is the version where you build a little every day until the wiring is automatic.
Common questions
Do I need to know how to code to be a cloud engineer?
You need to script, which is not the same as being a software engineer. If you can write a Python or Bash script that calls an API, loops over some resources, and fails safely when something is missing, you have enough. You do not need data structures interviews or a computer science degree. You do need to stop being afraid of a terminal.
Which cloud should I learn first, Azure or AWS?
Pick one and go deep. The fundamentals — networking, identity, compute, storage, IaC — transfer almost entirely; only the names change. Azure is the safe default if you are near Microsoft-heavy enterprises, government, or healthcare, which is most of the corporate job market. AWS is the default if you are aiming at startups. Learning one properly makes the second one a matter of weeks, not months.
How long does it take to learn the skills to become a cloud engineer?
Six to twelve months of consistent study for someone starting fresh, faster if you already work in IT support or as a sysadmin. The cert exam is the short part — a few weeks of recall. Becoming job-ready, meaning you have wired the skills together into something that deploys and survives, is the longer part, and it is the part employers screen for.
Can you become a cloud engineer without a degree?
Yes, and plenty do. Nobody senior in this field checks for a degree once you can show work. What replaces it is proof: infrastructure you built as code, a pipeline that deploys it, and the ability to talk through a decision you made and why. A degree helps get past some HR filters. Projects get you past the engineer who does the technical screen.