An applicant tracking system doesn't reject your CV for lacking keywords. It ranks and surfaces CVs, and a recruiter searches it. If your CV doesn't contain the words the recruiter searches for, you're not in the result set — and a CV nobody sees is functionally rejected. Keyword optimization is about being findable, not gaming a robot.

What the ATS actually does with keywords

Most modern systems don't auto-reject on a keyword score. They parse your CV into fields, store it, and let recruiters filter and search. The recruiter types "Kubernetes" or "financial modeling" or "contract law" and gets a ranked list. Your job is to make sure the words a recruiter would search for this role are present in your CV, in context, in the right place.

That reframe matters. You're not writing for an algorithm that counts. You're making sure a human's search returns you.

Step 1: Mine the job description, not the internet

The single best keyword source is the posting you're applying to. Generic "top ATS keywords" lists are useless because the recruiter is searching for that role's language, not the internet's.

• Paste the job description somewhere you can mark it up.
• Highlight every hard skill, tool, certification, and named methodology.
• Note the exact phrasing. If they wrote "customer success," don't only write "account management." If they wrote "CI/CD," include both "CI/CD" and "continuous integration" — recruiters search for either.
• Prioritize terms that appear in the requirements section and the job title. Those are what gets searched.

Step 2: Match the exact phrase and the common variant

Search is literal. "Project management" and "project manager" are different strings. A system may not connect "PM" to "product manager" unless both appear. So:

• Use the exact term from the posting at least once.
• Also include the obvious variant or expansion (acronym and full form).
• Don't rely on the system to infer synonyms. Some do; you can't tell which, so don't bet your application on it.

Step 3: Put keywords where they carry weight

Location matters for both parsing and recruiter trust:

• Skills section: the highest-density legitimate place for hard skills and tools. This is where a recruiter expects to scan them.
• Work experience bullets: keywords inside a real accomplishment ("migrated the billing system to AWS, cutting infra cost 30%") are worth far more than a bare list, because they survive the recruiter's read after they survive the search.
• A professional summary line: two or three core terms in the opening summary, naturally written.

The same keyword appearing in a real result and in the skills list is the pattern that works: findable by search, credible on read.

The stuffing tricks that backfire

Every shortcut here is known to recruiters and many modern parsers:

• White text or hidden keywords. Pasting keywords in white font or behind an image. Parsers extract text regardless of color, so the recruiter sees the dump in plain view. This reads as dishonest and ends the application.
• A wall of keywords with no context. A 40-term skills block with no evidence anywhere passes the search and fails the human in the same five seconds.
• Keywords for skills you don't have. It surfaces you for a search you can't survive in the interview. Wasting an interview slot on a mismatch burns a bridge.
• Stuffing the same term ten times. Repetition past two or three uses doesn't improve ranking in most systems and visibly degrades the writing.

A simple test before you submit

For the role you're targeting, ask: if a recruiter searched the three most obvious terms from this posting, would my CV come back, and would it survive their read once it did? If yes to both, you've optimized correctly. If you only pass the first, you've stuffed.

The honest summary

Mirror the posting's actual language, place those terms inside real accomplishments and a clean skills section, and cover the acronym-plus-full-form variants. That's the whole game. Everything beyond that — hidden text, keyword walls, claimed skills you lack — wins the search and loses the human ten seconds later, which is a worse outcome than never being found at all.