Cloud Career Guide: Job Roles and Responsibilities

In a company of fewer than two hundred people, there are rarely six cloud roles. There is one or perhaps two people who hold most of them, and a job title that undersells the range. This guide describes the six distinct responsibilities you will see advertised, what each one actually owns on a Tuesday afternoon, and which published figures are worth believing when you negotiate. The short version: build depth in one platform, learn cost discipline early, and stop treating architect as a job you get before you have operated anything.

The roles that actually exist in a small company

Job adverts invent titles faster than the industry invents work. Strip the marketing away and six distinct responsibilities remain.

• Cloud engineer. Builds and runs the infrastructure. Networks, compute, databases, identity, backups. The generalist role, and the one most small companies actually hire.
• Platform engineer. Builds the paved road other developers deploy on: pipelines, templates, environments, guardrails. Their customer is the internal development team, not the end user.
• Site reliability engineer. Owns availability as a measurable target. Service level objectives, error budgets, on-call rotations, incident reviews. Often the same person as the platform engineer in a small company, wearing a different hat on a bad week.
• Cloud architect. Makes the decisions that are expensive to reverse: which services, which regions, which data model, what the exit looks like. Part-time in most organisations, and badly done when the person has never been on call.
• Security engineer. Identity, least privilege, secret management, network boundaries, audit. Increasingly a separate hire once compliance questionnaires start arriving from customers.
• FinOps analyst. Owns the relationship between the bill and the business. Allocation, forecasting, commitment purchases, unit economics. The newest of the six and the one most often absorbed into a finance role.

The honest small-company picture: one cloud engineer doing four of these, an architect decision made once a year by whoever is most senior, and the finance manager quietly becoming the FinOps function because somebody has to explain the invoice. If that is your situation, it is not dysfunction. It is the normal shape of a thirty-person business, and it is a very good place to learn quickly.

What the salary data says, and what it hides

Salary data for cloud roles is noisy because job titles are inconsistent and self-reported surveys skew towards people who answer surveys. Two sources are worth quoting because they publish method and scale.

Three things that chart hides. First, it is management pay, which is the cleanest cross-country comparison the survey offers, not a cloud engineer salary. Second, Stack Overflow’s own data shows architects, founders and product managers earning less than engineering managers despite similar experience — roughly twenty years on average — which should temper the assumption that architect is the obvious promotion. Third, the figures are total compensation in local purchasing power, so a USD 52,000 median in India is not a quarter of the lifestyle of USD 200,000 in San Francisco.

The Dice 2025 report adds a useful detail for anyone planning the next two years: professionals designing, developing or implementing AI solutions reported salaries 17.7% higher than peers not involved in AI work. Dice also noted average tech pay has been roughly flat in real terms over nearly two decades once adjusted for purchasing power, which is a more sober frame than most careers content offers.

Cloud engineer, platform engineer or SRE

These three titles overlap enough that candidates and hiring managers frequently mean different things by them. A practical way to tell them apart is to ask who complains when the job is done badly.

If you are early in your career, aim at cloud engineer and get genuinely good at networking, identity and infrastructure as code. Those three are the foundations of all the others, and they are the skills that survive a change of provider or a change of fashion.

The FinOps job nobody budgeted for

The newest role on the list exists because of a measurement problem. Flexera’s 2026 State of the Cloud Report, drawn from 753 cloud decision-makers, found that 64% of organisations now measure the value cloud delivers to business units, up twelve percentage points year on year, while traditional cost efficiency metrics dropped six points. Nearly half, 49%, use unit economics to tie cloud cost to business outcomes, up from 40%. At the same time, wasted cloud spend rose to 29% after five years of decline, with AI workloads named as the cause.

Those two findings together describe a job. Somebody has to allocate spend to teams and products, forecast it, decide which commitments to buy, and answer the question nobody in engineering can: is this workload worth what it costs? In large organisations that becomes a FinOps practice. In a fifty-person business it is a day a month for whoever can read both a cost report and a product roadmap.

How roles map to certifications

Certifications are a hiring filter and a study structure. They are not evidence of competence, and experienced interviewers do not treat them as such. Used well, they sequence your learning.

The detail worth noticing is the recommended experience. Google suggests three or more years in industry, including a year designing and managing solutions on its platform, before attempting the professional architecture exam, and 20-30% of that exam is case studies. That is a sensible signal about the order of operations: operate things, then architect them.

Budget for renewals before you collect credentials. A two-year validity with a USD 100 renewal exam is manageable for one certification and becomes an annual subscription across three platforms. Go deep on one, stay conversant in the others.

Choosing a direction without guessing

A reasonable way to choose, in order of how much it affects your outcome.

1. Decide where you will work. Remote or relocated, this moves your pay band more than any skill decision. Stack Overflow’s 2025 data showed 45% of US developers working remotely, so the band is no longer purely a function of your postcode.
2. Pick one platform and go deep. Breadth across three providers reads as shallow. Depth in one plus fluent concepts reads as senior.
3. Get production experience, however small. One service you have operated through a real failure is worth more than a portfolio of tutorials.
4. Add one scarce skill. Cost engineering, security, or the AI-adjacent work Dice measured at a 17.7% premium. One, not three.
5. Write things down in public. A short, honest write-up of an incident you handled is the most persuasive artefact in the field, and almost nobody produces one.

If you are on the other side of this — a business trying to work out whether to hire, train or outsource these responsibilities — that is a different and equally common question. Our cloud solutions work is delivered remotely and frequently covers the gap while somebody is recruited. For context on the technical ground these roles cover, getting started in cloud technology sets out the learning path, our comparison of AWS, Azure and Huawei Cloud covers the platform decision, and our guide to software as a service explains the layer most of these roles end up administering.

Deciding whether to hire a cloud engineer or outsource the function for a year? Get in touch with Eudora Technology to talk about your project.