Aws Platform Engineer

An AWS platform engineer builds and looks after the cloud infrastructure that a company's software runs on, using Amazon Web Services (AWS - Amazon's rented computing platform, where servers, storage and databases are provisioned on demand rather than owned). Rather than writing the app itself, you provide the paved road the developers deploy onto: the accounts, networks, deployment pipelines, monitoring and security controls. Your "customers" are usually internal software teams, though you'll also answer to security, finance and whoever owns the cloud bill.

Approximate graduate salary

Typically somewhere around GBP 28,000-40,000 to start, though this varies widely. London and roles in finance or specialist consultancies sit at the upper end or above it; regional employers, smaller firms and the public sector usually sit lower. Some employers add on-call allowances or certification bonuses.

What you'd actually do

Writing and reviewing infrastructure-as-code - definitions of servers, networks and databases written in files (commonly Terraform or AWS's own CloudFormation/CDK) so environments can be rebuilt identically instead of clicked together by hand
Building and fixing CI/CD pipelines - the automated chain that takes a developer's code change, tests it and deploys it - so a merge results in a working deployment without anyone logging into a server
Responding to alerts and investigating why something is slow, failing or unexpectedly expensive: reading logs and metrics in CloudWatch or a tool like Datadog, Grafana or Splunk, and tracing the problem back to a config change, a resource limit or a bad deploy
Reviewing pull requests from other engineers and pairing with application developers who are stuck on permissions, networking or container config - a lot of the job is unblocking other people
Tightening access and security: writing IAM policies (the rules controlling which people and services can do what in AWS), rotating secrets, patching base images, closing off things flagged by security scanning tools
Cost work - spotting the forgotten test environment or oversized database that's inflating the monthly bill, and rightsizing or automating shutdowns
Writing documentation and runbooks, and taking part in incident reviews after something breaks, where the team works out what to change so it doesn't recur

How graduates get in

Graduate schemes at large employers - banks, insurers, telecoms, retailers, media, consultancies and the public sector all run technology graduate programmes with cloud, platform, DevOps or infrastructure streams. Common route, and usually the most structured.
Joining as a graduate software engineer and moving sideways into platform work after a year or two. Very common - many platform engineers didn't start there, and some employers prefer people who've felt the pain of deploying software before they build tooling for it.
Direct entry as a junior cloud/DevOps/platform engineer at a smaller company, scale-up or managed service provider. Fewer formal application windows, more emphasis on what you can show you've built.
Cloud and IT consultancies and AWS partner firms - these hire in volume and often put graduates through AWS certifications quickly, because certifications count towards the firm's partner status. A strong route if you want breadth across many client environments.
Degree apprenticeship or school-leaver route into IT infrastructure, then progressing into cloud work - not a graduate route as such, but it means you'll be working alongside people with a different background to yours.
Career-changer route via a conversion MSc in computing or a bootcamp, plus a home lab or personal projects. Possible but harder for platform roles specifically, because employers are less used to seeing infrastructure experience in bootcamp portfolios than web development.

What employers ask for

A degree is usually expected but the subject matters less than you'd think. Computer science, software engineering, maths, physics and other numerate or technical degrees are the most common; plenty of platform engineers come from unrelated subjects with demonstrable self-taught skills.
Many graduate schemes ask for a 2:1, some accept a 2:2, and a growing number have dropped degree classification filters entirely in favour of tests and practical exercises. This varies a lot by employer.
Evidence you can code, at least at scripting level. Python, Go and Bash are the usual currencies. You don't need to be a specialist developer, but 'I've never written code' is a genuine barrier.
Some hands-on AWS exposure - a personal project, university coursework, a placement year, or a free-tier account where you've actually deployed something. Employers care more about this than about certificates.
AWS certifications (Cloud Practitioner as a starter, Solutions Architect Associate as the common next step) help you get shortlisted, especially at consultancies, but rarely substitute for practical evidence. Some employers pay for and expect these once you're in.
A few roles - defence, government, some financial services - require UK security clearance or eligibility for it, which can mean residency requirements. Check job adverts carefully.

Skills that matter

Linux fundamentals

Almost everything you deploy runs on Linux, so filesystems, processes, permissions and networking commands are the vocabulary you debug in.

Scripting and automation mindset

The point of the job is that manual steps get replaced by repeatable code, so you're constantly asking 'how do I make this happen without a human doing it'.

Networking basics (DNS, IP ranges, load balancing, firewalls)

A large share of cloud problems are really networking problems wearing a costume - traffic not reaching the right place, or reaching somewhere it shouldn't.

Systematic debugging under pressure

When production is down you need to form and test hypotheses in order rather than randomly changing settings, and stay calm while people are watching.

Reading and writing clear technical documentation

Platform teams serve other engineers, and a well-written runbook or self-service guide prevents dozens of interruptions.

Security awareness

Misconfigured permissions and exposed storage are among the most common causes of real-world breaches, and it's your configuration that decides.

Cost consciousness

Cloud resources bill by the hour, so design choices you make show up directly on a finance report someone will ask you about.

Where it leads

Graduate or junior platform engineer: working on well-defined tickets with review, learning the estate, shadowing on-call before joining a rota.
Platform/cloud/DevOps engineer: owning services and pipelines end to end, on-call in your own right, leading small pieces of migration or tooling work. Most people reach this within a couple of years, but it varies with the employer and how much you're exposed to.
Senior engineer: designing how systems are built rather than just building them, setting standards, mentoring, and being the person consulted when something unusual breaks.
From there the paths fork. Technical: principal engineer, cloud architect, site reliability engineering, security engineering, or specialising in Kubernetes or data platforms. Managerial: platform team lead, engineering manager, head of infrastructure. Independent: contracting and freelance consultancy, which is common in this field and often well paid but without the security of permanent employment.
Timelines beyond the first couple of years vary enormously and depend far more on the scale of systems you get to work on than on years served.

What people get wrong

“It's basically IT support for the cloud - clicking around a web console and resetting things.”

Modern platform work is a software engineering job. Most changes are made by writing code in a repository, reviewed by colleagues and applied automatically; clicking in the AWS console to change production is often actively discouraged because it leaves no record.

“You need to collect AWS certifications to get hired and progress.”

Certifications help you get past a first filter and some employers require them commercially, but they test recall of AWS service names and options. Interviews tend to focus on whether you can debug a broken system and explain a design trade-off, which certificates don't demonstrate.

“You'll only ever need to know AWS.”

The job title says AWS, but the underlying skills - Linux, networking, containers, Terraform, pipelines, monitoring - transfer across providers, and many organisations run more than one cloud or still have on-premises systems you'll have to connect to.

“On-call means being woken up constantly.”

Most graduate roles ease you in with shadowing first, and a well-run team treats frequent night-time pages as a bug to be fixed rather than normal. That said, out-of-hours responsibility is a real feature of the job at many employers and you should ask about the rota, compensation and alert volume at interview - practice differs enormously.

Where this varies

The title covers quite different jobs. At a large bank or government body you may work on one narrow slice of a heavily governed estate, with change approval processes and long lead times. At a scale-up you might own the whole platform with far more autonomy and far more chaos. At a consultancy or AWS partner you'll move between client environments every few months, which builds breadth fast but gives you less chance to live with the consequences of your own designs. Job titles are used loosely too - 'platform engineer', 'DevOps engineer', 'cloud engineer', 'infrastructure engineer' and 'site reliability engineer' often describe overlapping work, so read the advert rather than the title. Hybrid working is common in this field, and fully remote roles exist, but this differs by employer and by whether the systems you touch are security-restricted.

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.