From survival job to cloud engineer: the path nobody glamorizes

You are working the warehouse, the night shift on the ward, the airport security post, the rideshare app — and studying cloud on the train home. This is the honest arithmetic of that path: what to learn in what order, how long it really takes at 8–10 hours a week, and when it is actually safe to quit.

New to cloud? CAMPUX is a free, build-first course. Start here →

Going from a survival job to a cloud engineering career realistically takes 12 to 18 months at 8 to 10 study hours a week around full-time shifts: fundamentals first, then Azure administrator skills, then one real public project — and no honest version of this fits in three months.

I want to be straight with you from the first paragraph, because you have already seen the other version of this article. The one with a stock photo of a man smiling at a laptop, promising a six-figure cloud job in 90 days. You have probably also seen someone from your community WhatsApp group pay for that promise and get nothing. So this piece will do the opposite: it will give you the math, the order of operations, the part where the plan usually breaks, and the one decision — when to quit the survival job — that people get wrong in both directions.

A word on who is writing this. I taught myself into this field. No computer science degree, no tech pedigree — I was a helicopter pilot before I was an engineer. Nine years in IT later, six of them on Azure, I work as a Senior Cloud Engineer and Cloud InfoSec Program Manager. I am not telling you the door is easy. I am telling you it is real, because I walked through it from the outside, and the people I now watch walk through it fastest are the ones with exactly your constraints.

First, respect the job you already have

Nobody glamorizes the survival job, so let me at least be accurate about it. If you are doing home-health shifts, warehouse picking, overnight security, or driving after your day job — while sending money home every month — you are already running a schedule and a budget that would break most of the people writing "learn cloud fast" articles. That is not a consolation prize. It is the single most predictive trait for finishing this path.

Here is the thing I have come to believe watching career-changers: the comfortable ones fail more. Someone with a quiet desk job and free evenings has weak reasons to study on a Tuesday night, and weak reasons lose to the couch. You do not have that problem. Your reasons are sitting at your dinner table and waiting on your remittance. When a person like that puts in 90 focused minutes after a 12-hour shift, those minutes are worth more than three hours of a bored person's browsing. The constraint you think disqualifies you — fragmented time, real stakes — is the discipline engine the whole plan runs on.

And your work history is not the blank slate American hiring makes it feel like. If you ran a banking operations desk in Lagos, you already understand access control and audit trails — that is half of cloud security thinking. If you were a nurse, you understand checklists, escalation, and what happens when a handoff fails — that is incident response. The technical stack is new. The judgment underneath it is not. I wrote more about translating that history in cloud jobs with a foreign degree, and if you are still deciding whether cloud is even the right target, start with the map of cloud computing for immigrants in the US.

The honest arithmetic: 8–10 hours a week, 12–18 months

Let us do the math that the 90-day ads refuse to do. Becoming employable in cloud — not certified, employable — takes somewhere in the region of 400 to 600 hours of real work: learning fundamentals, building administrator-level skills, doing labs that break and get fixed, and producing one project you can show a stranger. A full-time student can compress that into six months. You are not a full-time student.

Your realistic budget looks like this: two or three 90-minute blocks on weeknights — after shifts, on the train, at 5am before the house wakes — plus one protected Saturday or Sunday morning block of three to four hours. That is 8 to 10 hours in a good week. Some weeks will be worse: a double shift, a sick child, Ramadan or a funeral back home. Build the plan on 8 hours, and the bad weeks bend the timeline instead of breaking it.

At that pace, 400–600 hours is roughly 50 to 65 weeks. Call it 12 to 18 months, including a job search that overlaps the final months. Here is how the phases lay out:

Treat the calendar numbers as a range, not a promise — people with prior IT exposure land at the front of it, people who lose whole months to life land past it. But the direction is settled: this is a one-to-one-and-a-half-year project, and anyone selling you a 90-day version is selling to your desperation, not your future. I went deeper on the time question in how long it takes to learn Azure.

Notice also what the cost column says. The study material itself can be free — ours is, all 42 classes of it — and the certifications together cost less than one week of most survival jobs. The scarce resource in this plan is never money. It is your protected hours, and whether the path you follow wastes them.

What to study, in what order

Phase 1: fundamentals — learn the map before the streets

Start with what cloud actually is: someone else's data centers, rented by the hour, controlled through code and consoles. You need the vocabulary — compute, storage, networking, identity — and the mental model of how a company's systems live in Azure. This is AZ-900 territory. Do not rush past it; every confusing thing in phase 2 is a fundamentals gap wearing a disguise. If you are starting from zero, begin at Class 1: what is cloud and just keep walking forward.

Phase 2: administrator skills — the AZ-104 layer

This is the longest phase and the employable core: creating and securing virtual machines, wiring virtual networks, managing storage, running identity and access through Entra ID and RBAC, watching costs and monitoring. The certification matters less than the hands doing the work — an interviewer can tell in four minutes whether you have actually built a VNet or only watched one being built. Do the labs. Break things. When your role assignment fails at 1am and you figure out why, that hour is the education.

Phase 3: one real public project — your American work history

You fear the "no US experience" catch-22, and the answer to it is a public project: something deployed to Azure, defined as infrastructure code, documented in a README, sitting in your GitHub where any hiring manager can inspect it. One good project beats five certificates, because it is the one thing on your résumé nobody has to take on faith. This is how you manufacture experience when no one will give you any — I laid out the mechanics in how to get Azure experience without a job.

Phase 4: the search — run it from a position of strength

US hiring has mechanics nobody back home teaches: one-page résumés with no photo and no age, ATS keyword filters that reject you before a human looks, and referrals that move applications from a pile of 400 to a pile of 12. Your community is a real network here — the cousin in Texas "in IT," the association WhatsApp group — use it without embarrassment; referrals are how Americans get jobs too. And on the sponsorship question: if you hold a green card or an EAD, the answer to "will you now or in the future require sponsorship?" is no, and that single answer removes a filter that screens out a large share of applicants. The details — and what the question actually means — are in Azure cloud engineer jobs and visa sponsorship; for anything specific to your status, confirm with your attorney. The job-search craft itself is Class 42: landing the job.

Certificates say you studied. A deployed project says you can work. American hiring managers can only see the second one.

The honest gap: certified is not hired

Here is the section the course-sellers cannot write, because their product ends at the certificate. Passing AZ-104 does not get you a job. I have watched people collect three certifications and twelve months of rejections, and the pattern is almost always the same: they optimized for passing exams, because exams give a date, a score, and a dopamine hit — and job-readiness gives you none of those. The market is full of certified people. It is short on people who can sit in a screen and walk through a real system they built, explain why they chose what they chose, and reason out loud when asked something they do not know.

That is why phase 3 is not optional and not a garnish. The project is the bridge across the certified-to-hired gap, and it is also your armor in the interview: when your accent makes you self-conscious on a phone screen — and I know that fear is real — the conversation going well depends far less on your vowels than on whether you have concrete work to talk about. Confidence in interviews is mostly just having done the thing. On salary, I will keep to the discipline I am preaching: US cloud engineer postings commonly show ranges from around the high-$80,000s to the $130,000s on Glassdoor and Levels.fyi as of 2026, varying hard by city and level. Treat any specific number with suspicion; treat the gap between that and a survival wage as settled.

The trap: quitting the survival job too early

Now the decision that ruins more of these journeys than any technical failure. Somewhere around month six, when the studying is going well, a voice will say: imagine how fast I could go full-time. Sometimes the voice is a program selling you that plan. Run the numbers before you listen.

Quitting turns a time problem into a money problem, and the money problem has a shorter fuse. Once you are ready to interview, a first cloud job commonly takes three to six months of applications and screens — longer for a first role than for an experienced hire, and your savings do not pause for it. Neither do the remittances; the family in Kumasi does not stop counting on you because you enrolled in something. And the cruelest part: financial desperation leaks into interviews. A candidate who needs this exact offer to make rent negotiates worse and reads as anxious. The candidate with a paycheck reads as solid, and can say no to the lowball.

The survival job, for all it takes out of your body, is the thing funding the transition. It is not the enemy of the plan. It is the floor under it.

When it is actually safe to jump

The clean rule costs nothing extra, because this path was built around shifts from the start: you leave when you have a signed offer with a start date. Not a verbal maybe, not a recruiter's enthusiasm — a signed offer. Since you never had to quit to study, you never have to gamble.

If life forces an earlier exit — the body giving out, a layoff, a schedule that makes study truly impossible — then the minimum bar for a deliberate jump is all four of these, not two: savings covering at least six months of real expenses including what you send home; the project finished and public; the résumé and interview prep done, not planned; and applications already producing callbacks, because callbacks are the only market signal that counts. Anything less, and you are not jumping — you are falling with a plan.

There is one honorable half-step worth naming: some people bridge through an IT support or help-desk role — lower pay than the eventual target, but IT experience on a US résumé while you finish the cloud skills. It lengthens the total path a little and de-risks it a lot. No shame in a bridge.

Questions people also ask

How long does it take to go from a survival job to a cloud job?

Plan on 12 to 18 months if you are studying 8 to 10 hours a week around full-time shifts. That covers cloud fundamentals, Azure administrator skills, one real public project, and a realistic job search of three to six months. Some people move faster with prior IT background or more study hours; almost nobody working full-time does it in three months, whatever the ads say.

Can I become a cloud engineer while working full-time night shifts?

Yes, and most people in your position do exactly that. The path is built on fragmented time: 90-minute study blocks after shifts, one protected weekend morning, and labs you can pause and resume. It takes longer than studying full-time — 12 to 18 months instead of 6 to 9 — but you keep your income the whole way, which is what makes the plan survivable. Consistency beats intensity here.

Should I quit my job to study cloud computing full-time?

For most people supporting a family, no. Quitting compresses the study timeline but starts a financial clock that runs out faster than the job search does — a first cloud job commonly takes three to six months of applications after you are ready, and desperation shows in interviews. Keep the survival job, study around it, and only jump once you have an offer in hand or savings that cover 6 or more months plus a finished portfolio.

Do I need a degree to become a cloud engineer in the US?

No. Many cloud job postings list a degree as preferred rather than required, and demonstrated skills — certifications plus a real project a hiring manager can inspect — carry more weight than a credential for entry-level cloud work. If you hold a foreign degree, a credential evaluation from a service like WES can make it legible to US employers, but do not wait on that to start studying. The portfolio matters more.

When is it safe to leave my survival job for tech?

The clean answer: when you have a signed offer with a start date, and not before. The offer-in-hand rule costs you nothing — cloud studying fits around shifts precisely so you never have to gamble your rent on a job search. If you must leave earlier, the minimum bar is savings covering at least six months of real expenses, a finished public project, an interview-ready résumé, and applications already producing callbacks.