Hiring guide
How to Write a Software Engineer CV That Wins Interviews
Your CV now has two readers, and they are looking for different things. The first is a machine that decides in under a second whether you are even seen. The second, if you get past the first, is a human, and on our desk that human is a former engineer. Write for both, in that order.
I read CVs for a living now, having spent years writing production code, so this is the guide I wish more engineers had. None of it is about tricks. It is about making genuine signal easy to find.
First, get parsed
Modern applicant tracking systems screen with natural-language processing and machine learning, compare you against the job description, and rank you. A human recruiter might give a CV a couple of minutes; the software can reject it in less than a second if it cannot read it.
So make it readable. Use a clean single-column layout with standard headings, Summary, Skills, Experience, Education, a normal font and plain bullet points. Avoid the things that parse badly: tables, multiple columns, graphics, icons and any text trapped inside an image. Send a PDF unless the portal asks otherwise, as it holds layout while staying machine-readable. Match the real language of the job, but do not stuff keywords. Screeners now weight evidence over repetition, and buzzwords without supporting bullets can count against you.
Then win the human
Once a person reads it, the single biggest lever is this: write impact, not responsibilities. “Responsible for the payments service” tells me nothing. “Rebuilt the payments service to cut checkout latency from 900ms to 180ms, serving 50,000 daily users” tells me the scale you operate at, that you owned the decision, and what changed because of you. Those three things, scale, ownership and measurable outcome, are exactly what an experienced reviewer scans for.
You do not need a metric on every line, and inventing them is worse than omitting them. But your best two or three bullets per role should carry a number and a decision you made, not a task you were assigned.
Know what actually differentiates in 2026
Something real has shifted. Programming languages and frameworks have moved from the thing that gets you hired to table stakes: still necessary, no longer decisive. What separates candidates now is architecture and judgement, how you work with AI tooling, and code-quality discipline. The strongest CVs I see show someone directing AI and applying engineering judgement to it, catching what it gets wrong, rather than simply shipping faster. Reasoning and taste are the scarce signals; raw output is not.
This mirrors what we test for when we interview. Our guide on how to interview backend engineers covers the other side of the same coin.
Beat the AI-slop problem by being specific
Hiring inboxes in 2026 are full of generic CVs auto-generated to match a spec, and reviewers have learned to skim straight past them. Using AI to tidy your structure is fine. Using it to manufacture substance is what gets you ignored, because it produces exactly the bland, interchangeable prose everyone now recognises. Your advantage is the specificity only you can supply: the actual numbers, the real trade-off you weighed, the thing you genuinely built. That is unfakeable, and it is what a former engineer is reading for.
The practical checklist
Keep it to one page under ten years of experience, two beyond that, and only if every line adds value. Open with a two-to-four-line summary stating your role, years, domains and a proof point or two, and cut the fluffy adjectives. Link a GitHub or portfolio only if it shows real work. Put the most relevant role and your strongest results first. Then delete anything that is a duty rather than an achievement.
Where this leaves you
If your CV clears the machine and shows genuine signal, the interview is yours to win. It is worth benchmarking your experience against the market too: our UK tech salary statistics show what different levels and skills actually pay in 2026, so you can pitch yourself accurately.
This is precisely how we read CVs. Every engineer we put forward is assessed by people who have done the job, across the industries we hire for, so genuine signal is rewarded and polish alone is not. If you are looking for your next role, see our open engineering roles, find out how we work with the clients we place engineers into, or talk to us and we will represent you properly.
FAQ
Frequently asked questions
What should a software engineer CV include in 2026?
A short summary (role, years, domains, one or two proof points), a skills section that matches the job, and experience written as impact rather than duties. For each role, show the scale you worked at, the technical decisions you owned, and the measurable outcome. Keep it to one page under ten years of experience, two pages beyond that, and only if every line earns its place.
How do I get my CV past ATS and AI screening?
Use a clean single-column layout with standard headings (Summary, Skills, Experience, Education), a readable font and normal bullet points. Avoid tables, columns, graphics and text in images, which parse badly. Submit a PDF unless told otherwise. Match the real language of the job description, but do not keyword-stuff: modern screeners weight evidence over repeated buzzwords and can penalise stuffing.
Do programming languages still matter on a CV?
They are necessary but no longer what gets you hired. In 2026 the differentiators are architecture and technical judgement, how you work with AI tools, and code-quality discipline. List the languages and frameworks the role needs, then spend your best lines on decisions you made and problems you solved, not a long inventory of technologies.
Should I use AI to write my CV?
Use it to tidy structure and phrasing, but not to generate the substance. Recruiters and hiring managers are now flooded with generic AI-written CVs and have learned to skim past them. Your edge is specificity only you can supply: the real numbers, the real trade-offs, the real thing you shipped. Showing you direct AI and catch what it gets wrong is itself a strong 2026 signal.