Platform Engineer

A platform engineer builds and maintains the shared infrastructure and internal tooling that software development teams use to build, test, deploy and run their applications. Rather than writing the features customers see, you provide the "paved road" - cloud environments, deployment pipelines, monitoring, security guardrails - so that other engineers can ship code quickly and safely. Your customers are usually the developers inside your own organisation.

Approximate graduate salary

Roughly GBP 28,000-40,000 to start as an approximation, and this varies a great deal - London and large financial or tech employers sit at the top of that range or above, while regional employers, smaller companies and public sector roles typically sit lower. Pay in this field tends to rise quickly with a few years of experience.

What you'd actually do

Writing and reviewing infrastructure-as-code - configuration files (commonly Terraform, or cloud-native equivalents) that define cloud resources such as servers, databases, networks and permissions, so environments can be rebuilt reliably rather than clicked together by hand
Building and fixing CI/CD pipelines - the automated sequence that takes a developer's code change, runs tests on it and deploys it to live systems - and working out why a build has started failing
Working with containers and orchestration: packaging applications (typically with Docker) and running them on Kubernetes or a managed cloud container service, including tuning how much memory and CPU they get
Responding to alerts and investigating incidents when something is slow or down - reading logs, dashboards and traces to work out whether the problem is the application, the infrastructure or a third-party dependency, then writing up what happened afterwards
Scripting and small-scale automation in Python, Go or Bash to remove repetitive manual work, such as rotating credentials, provisioning accounts or cleaning up unused cloud resources
Helping development teams who are stuck - answering questions in a support channel, pairing on a deployment problem, or writing documentation and templates so the same question doesn't come up ten more times
Cost and security housekeeping: spotting cloud spend that has crept up, patching base images, tightening access permissions, and feeding into decisions about which tools the organisation should standardise on

How graduates get in

Technology or software engineering graduate schemes at large employers (banks, retailers, telecoms, media, consultancies, government departments) - many rotate you through development, infrastructure and operations before you specialise, and platform or 'cloud engineering' is a common landing spot. This is the most structured route.
Direct entry as a junior software engineer or junior developer, then moving sideways into platform work after a year or two once you've seen how deployment and infrastructure actually work. This is extremely common - many platform engineers did not start with that title.
Entry via IT support, service desk, network or systems administration roles, then learning cloud and automation on the job. This route is more common in organisations with a traditional IT function, and often doesn't require a degree at all.
Degree apprenticeships in digital and technology solutions, which combine paid work with a part-time degree - a growing route and a genuine alternative to the graduate scheme.
Industrial placements or summer internships in engineering teams during a computing degree, which frequently convert into graduate offers; relevant placement experience matters more to many employers than your exact module list.
Self-taught or conversion routes - a non-computing graduate with cloud certifications and a portfolio of personal projects can get in, though usually at smaller companies or via a junior/associate role rather than a competitive graduate scheme. Bootcamps exist but vary enormously in how well employers regard them.

What employers ask for

Most employers prefer a degree in computer science, software engineering or a closely related subject, but other numerate or technical degrees (maths, physics, engineering) are widely accepted if you can show coding ability. Some employers now accept any degree subject plus demonstrable skills.
A 2:1 is the common cut-off on structured graduate schemes, and some ask for specific A-level grades or UCAS points. Smaller companies and direct-entry junior roles often don't filter on grades at all and care more about what you can build.
Evidence that you can actually write code - a public code repository, university projects, a home lab, a deployed hobby project. For platform roles specifically, having set something up end to end (a small app running in the cloud with automated deployment) is worth more than listing technologies.
Cloud certifications from the major providers (associate-level cloud practitioner, solutions architect or developer certifications) are not usually required but are genuinely useful as a signal for career changers and self-taught candidates. Many employers will pay for them once you're in.
Familiarity with Linux and the command line, and with Git version control, is close to universally expected - more so than any particular cloud platform, which you'll be trained on.
Security clearance is required for some public sector, defence and critical-infrastructure roles, which usually means UK residency requirements - check this before applying as it can rule out otherwise suitable candidates.

Skills that matter

Debugging systems you didn't build

Most of the job is working out why something that used to work has stopped, across layers of code, network and cloud configuration that nobody fully documented.

Scripting and automation mindset

The value you add is turning a manual task someone does every week into something that runs itself reliably, which means noticing the repetition in the first place.

Understanding of networking and operating system fundamentals

A huge share of cloud problems come down to DNS, firewall rules, certificates or permissions, and you cannot bluff your way through those.

Treating internal developers as users

Platforms fail when they're technically elegant but painful to use, so you need to ask teams what slows them down and design tooling they'll actually adopt rather than route around.

Writing clear documentation and runbooks

You're building things other teams depend on at 2am, and a readable guide to 'what to do when this alert fires' is often more valuable than the code itself.

Staying calm and methodical under pressure

During a live incident with people watching, the useful engineer is the one working through hypotheses in order rather than changing three things at once.

Where it leads

Graduate or junior platform/cloud engineer: working on well-defined tickets with close review, learning the organisation's stack, usually shadowing on-call rather than carrying it alone.
Platform or DevOps engineer (typically after a couple of years, though this varies): owning parts of the platform, joining the on-call rota properly, leading smaller pieces of work end to end.
Senior platform engineer or site reliability engineer (SRE): designing how systems should work, setting reliability standards, mentoring juniors, and making trade-off decisions about cost, security and speed. Timelines here vary widely - some reach this in four or five years, others take longer, and it depends heavily on the size and pace of the employer.
From senior, the path commonly forks: a technical track towards principal engineer, cloud architect or security engineering; or a people track towards engineering manager and head of platform. Neither is better paid by default.
Longer term, the skills travel well into consultancy, contracting (day-rate work is common and can pay well but brings no job security), specialist cloud security, or technical roles at cloud and tooling vendors.

What people get wrong

“It's an IT support or systems administration job with a fancier name.”

Modern platform engineering is a software engineering discipline - you spend much of your time writing code and configuration that is version-controlled, tested and reviewed like any other codebase. If you dislike programming, you will probably dislike this role.

“You need to be an expert in Kubernetes and a long list of tools before anyone will hire you.”

Graduate employers expect fundamentals - Linux, networking, Git, one programming language - and train you on their specific toolchain, which will be different at your next employer anyway. Tool lists in job adverts are usually a wish list, not a checklist.

“Because it's infrastructure, it's a solitary, low-communication job.”

You're effectively running an internal service for other engineering teams, so a lot of the week is conversations, support requests, negotiating standards and persuading teams to adopt your tooling rather than build their own version.

“On-call means being woken up constantly.”

Practice varies enormously. Some organisations have a quiet, well-compensated rota you join only after months of support; others are genuinely disruptive, and a few have no out-of-hours expectation at all. It is a reasonable and normal thing to ask about at interview, including whether it's paid.

Where this varies

The job title is used inconsistently and overlaps heavily with "DevOps engineer", "cloud engineer", "infrastructure engineer" and "site reliability engineer" - read the advert rather than the title. At a large organisation with a dedicated platform team, you'll likely specialise in one area (pipelines, Kubernetes, observability, cloud security) and build self-service tooling for dozens of teams. At a small company, "platform engineer" often means you are the whole infrastructure function and also fix laptops. Sector matters too: financial services and government bring heavy compliance, change-approval processes and audit trails that slow things down deliberately; consultancies involve moving between client environments and more client-facing work; product companies tend to give more autonomy and faster release cycles. Remote and hybrid working is relatively common in this field compared with many graduate roles, though some regulated employers require more office presence.

General guidance about the role across the UK market, not about any specific employer. Entry routes and requirements vary — always check the individual job advert.