How to Write a CV: What ATS and Recruiters Actually See

You probably think your CV is being read. In practice, most CVs are parsed by software, filtered by algorithms, and then scanned by a human for under 10 seconds. Read is too generous a word for most of what happens to them.

How to write a CV that works in 2026 is not the same question it was five years ago. Your CV now has two audiences, and they evaluate it in completely different ways. The first is a system: an Applicant Tracking System that extracts, structures, and matches your information against the job description. The second is a human: a recruiter who will spend 6 to 10 seconds deciding whether to keep reading. The CVs that fail are almost never the ones with the wrong experience. They are the ones that make the right experience invisible to one audience or both.

What ATS Does (and Does Not Do)

Candidates often talk about ATS as if it is a judge that decides whether they are qualified. It is not. An ATS is a parser. It reads your document, extracts data into structured fields (name, contact info, work history, skills, education), and compares that structured data against the job description. If the match is strong enough, your application moves forward. If not, it gets deprioritized or filtered before a recruiter sees it.

The problem is that parsing is fragile. Greenhouse, one of the most widely used ATS platforms, documents that resume parsing can fail because of graphics, photos, tables, column layouts, headers and footers, text boxes, and files over 2.5 MB. According to their 2025 Workforce Report, 53% of US recruiters now say that ATS or AI systems complete their initial screening. A quarter of those recruiters are not confident in what their own systems prioritize.

What this means in practice: the system is not reading your CV the way you think it is. It is extracting fragments of text and trying to match them to a job profile. If the extraction fails because of how your document is formatted, the content does not matter. You are filtered out before anyone evaluates what you wrote.

The Format That Survives Both Audiences

The safest architecture for a CV in 2026 is conservative. Not because creativity is bad, but because the parsing layer does not reward it, and the human layer does not need it.

Use a single-column layout. Put your name, email, phone, location, LinkedIn, and portfolio in the document body near the top, not in the header or footer. Use standard section names: Summary, Skills, Experience, Education, Certifications. Save as a searchable PDF. That is the baseline.

But format is only half of it. When I opened a CV, I needed to answer four questions in under 10 seconds: what level are you, what kind of roles fit you, what is your strongest domain, and what outcomes have you delivered. If those answers are not visible immediately, the CV goes to the bottom of the stack. Not because I was careless. Because I had 400 more to get through.

The test I recommend is what I think of as the compression test. At every stage of the hiring process, your profile gets summarized. A recruiter describes you to a hiring manager in one sentence. A hiring manager pitches you to the team. Even when I was the one making the decision myself, I still needed to compress a candidate into a clear signal: "Senior frontend engineer, strong in React and performance, led a migration that improved conversion by 14%." If your CV does not naturally collapse into something like that, the document is still too vague.

Keywords Without Keyword Stuffing

ATS keyword matching is a real mechanism, and ignoring it is a mistake. But the way most candidates approach it creates a different problem.

The principle is straightforward: mirror the job description's exact language where it honestly describes your experience. If the role asks for "TypeScript, React, Next.js, GraphQL, AWS," those exact terms should appear in your CV if you have them. Do not rely on synonyms. "Project management" and "program management" might describe overlapping skill sets in practice, but to an ATS they are different strings.

Where candidates get into trouble is the other direction: padding the CV with terms they barely know, inflating skill lists, or hiding keywords in white text. LinkedIn's 2026 talent research found that 93% of talent acquisition professionals believe accurately assessing skills is crucial for quality of hire. The market is moving toward skills-based hiring, and that means both systems and humans are getting better at spotting mismatches between what a CV claims and what a candidate can actually demonstrate.

The CVs that listed 20 technologies at "expert" level were always the ones I trusted least. A CV with eight skills listed precisely, each one backed by evidence in the experience section, is more credible than one with forty skills and no proof. Trust is thin in this market. Precision builds it faster than breadth does.

The practical solution is a categorized skills section. Instead of a single block of keywords, group them: Languages, Frameworks, Cloud/Infrastructure, Data, Testing, Methods. This is easier for the ATS to parse and easier for a human to scan.

The Bullets That Make a Recruiter Stop Scrolling

Once the CV passes the system, it reaches a human. And the optimization problem changes entirely.

I was not looking for impressive. I was looking for evidence. When I read an experience bullet, I was asking: what did this person do, at what scale, with what tools, and what was the result? Bullets that answered all four questions made me slow down. Bullets that answered one or two made me keep scanning.

The first version tells me what you touched. The second tells me what you accomplished, with whom, using what, and why it mattered. Only the second gives me something I can forward to a hiring manager.

What I wish more candidates understood is that seniority is not communicated through adjectives. Words like "senior," "strategic," "lead," and "expert" are cheap, and every CV has them. What actually convinced me is evidence that you made tradeoffs, influenced direction, improved a system, or increased leverage for others. Architecture decisions, incident reduction, platform improvements, mentoring, cross-team alignment: those are the signals I look for when I am trying to figure out whether someone actually operates at the level they claim.

The Mistakes I See Every Week

These are the patterns that showed up repeatedly in my screening. They are not fatal on their own, but each one reduces your odds of making it through the first pass.

Beautiful templates that break everything. The CV looks stunning as a PDF. The ATS cannot read it. Tables, graphics, icons as skill bars, multi-column layouts: these are design choices that actively work against you. I have seen talented candidates get filtered out because their beautifully designed CV parsed as an empty profile.

Generic summaries. "Results-driven professional with a passion for excellence." That sentence tells me nothing. A strong summary defines your seniority, domain, core tools, and what makes you different, in three to five lines. If your summary could belong to anyone, it is not doing its job.

Inconsistent external signals. After reading a CV, I almost always checked LinkedIn. If your CV says "Senior Product Engineer focused on performance" but your LinkedIn headline says "Growth Hacker | AI Visionary | Full Stack Ninja," you have created doubt. The two documents should tell the same story.

Inflated skill lists. Twenty frameworks, ten cloud platforms, eight databases, all at "expert" level. This does not communicate range. It communicates that you are not being honest about depth. The candidates who list fewer skills with more evidence are the ones who earn my trust.

Unexplained gaps. An ATS does not judge employment gaps. Humans do not judge them either, when they are explained. What creates suspicion is ambiguity. If you took time for consulting, caregiving, study, or a career pivot, just say so. One line of context removes the question entirely. Silence forced me to guess, and guessing rarely favored the candidate. For a framework on how to address gaps in the interview itself, read how to explain a career gap.

The CV That Gets You Hired

The CV that works in 2026 is not the most creative one or the most detailed one. It is the one that makes your value obvious to a machine in milliseconds and to a human in seconds.

Clean format. Honest keywords. Evidence in every bullet. A summary that tells me who you are and what you bring. External signals that match what the document says. That is the entire formula, and it is not complicated. The candidates who follow it are the ones I call back. And they are the ones who, when they walk into the interview, have already started building credibility before they say a word.