How to Scan Your Resume Against Any Job Posting

If your resume isn't being read, it's almost always a keyword-matching problem. The fix is faster than rewriting the whole resume — it's a 30-second scan that surfaces what's missing. This post explains exactly how the scanning works, what to look for in the output, and how to act on it without lying.

What a resume keyword scanner actually does

A keyword scanner takes two inputs:

1. Your resume (DOCX, PDF, or pasted text)
2. A job description

…and produces:

1. A match score — typically 0-100%, sometimes called "ATS score" — indicating how well the resume matches the JD
2. A list of missing keywords — terms in the JD that don't appear in your resume
3. (Sometimes) A rewrite suggestion showing where to insert each missing keyword

Behind the scenes, modern scanners do two things:

Exact-phrase matching: They look for literal substring matches between your resume and the JD's vocabulary. If the JD says "Kubernetes" and your resume says "Kubernetes," that's a match. If the JD says "TypeScript" and your resume says "Typescript" (different capitalization), most scanners count that as a match too. Semantic similarity: Newer scanners (including ours) use embedding-based comparison. The model encodes the resume and the JD as vectors and computes a cosine similarity score. This catches near-misses — "container orchestration" vs "Kubernetes" — that exact-string matching doesn't.

The semantic part is important because real ATSes (Greenhouse, Lever, Ashby) increasingly use semantic scoring. A scanner that only does string matching will overstate your real ATS score. A scanner that does both gets you closer to what the actual ATS sees.

The two ATS architectures and why it matters

Not all ATSes score the same way:

Exact-match ATSes (older / commodity): Some employers run ATSes that just count keyword frequency. If the JD says "Python" 3 times and your resume says "Python" 1 time, you score lower than someone with 4 mentions. This rewards keyword stuffing. Semantic-match ATSes (modern): Greenhouse, Lever, Ashby, and the Workday-2024-onward versions use semantic similarity. Stuffing "Python python python3 Python 3.x" actively hurts you (looks unnatural). Using "Python" once in genuine context within a strong bullet rewards you more.

A good scanner tells you which architecture the company is likely on. Most do this by looking up the company's career page and identifying the ATS subdomain (greenhouse.io, lever.co, ashbyhq.com, etc.).

What good scanner output looks like

A useful scanner output has at least these elements. Anything less and the tool isn't worth paying for:

1. Numeric match score with a target threshold ("aim for 75%+")
2. Missing hard skills — concrete tools, frameworks, technologies in the JD that aren't in your resume
3. Missing soft skills / phrases — things like "cross-functional collaboration" or "stakeholder management" that may matter
4. Format flags — issues like missing contact info, two-column layouts, embedded images that the ATS might fail to parse
5. A specific list of bullet rewrites showing where to insert missing keywords naturally

A bad scanner just gives you a score and a list. You're left to figure out where to put the keywords yourself, which is the actual hard part.

How to use a scanner without lying

The risk with keyword scanners is that they incentivize keyword-stuffing — adding terms that aren't actually true skills. This will get you to interview, then immediately fail.

The right approach:

1. Run the scan.
2. For each missing keyword, ask yourself: have I done this thing, even tangentially?
3. If yes, add the keyword to a bullet that's a true description of work you actually did.
4. If no — don't add it. Skip and move on.

Example. JD says "Kubernetes." Your resume doesn't mention K8s. Have you ever written a Helm chart, deployed via Argo, or maintained kubectl scripts? If yes, add a bullet using the K8s keyword in real context. If you've literally never touched K8s — don't add it. The interview will catch you.

The right outcome of a keyword scan is "my resume now reflects what I've actually done, in the language the JD uses." The wrong outcome is "I added 8 buzzwords I can't defend in an interview."

How to scan against any job posting in 30 seconds

The specific workflow that takes 30 seconds:

1. Paste the JD URL into our keyword scanner, or copy-paste the JD text.
2. Upload your resume (DOCX or PDF — DOCX preferred for parsing).
3. Read the output: match score, missing keywords by category, format flags.
4. For each missing keyword you've actually done: add it to the most natural bullet.
5. Re-scan. The score should jump 15-30 points.

Total time: 30 seconds for the scan, 5-10 minutes for the rewrites.

If you're applying to 10 jobs, this is your single highest-leverage action. Tailoring takes <10 minutes per JD and lifts response rate noticeably.

What good keyword density looks like

A common question: "how many times should I include the keyword?"

Best answer based on testing across major ATSes in 2026:

• Hard skills (Python, AWS, Kubernetes, Salesforce): Once in your skills section + once in a relevant bullet = 2 mentions. More than 4 looks stuffed.
• Job titles: 1-2 mentions is fine. Don't repeat your own title in every bullet.
• Soft skills phrases: Once is enough. Recruiters find repeated soft-skill phrases ("strong communication skills... excellent communication... communication...") immediately suspicious.
• Buzzwords (synergy, results-driven, dynamic): Zero is correct.

The pattern: mention each important keyword 1-3 times, in real context, distributed across sections. Not 8 times in one paragraph.

When scanners are wrong

A few cases where the scanner score doesn't predict real outcomes:

1. Company runs exact-match ATS, but the scanner uses semantic. You'll show 80% match, but the company's older Workday parses your resume by counting exact strings — you may score lower in reality.
2. Industry-specific terminology drift. Healthcare, government, defense, finance — these have specialized vocabularies where scanners miss subtle matches. A nursing scanner needs to know that "Med-Surg" and "Medical-Surgical" are the same thing. A generic scanner doesn't.
3. The JD itself is keyword-stuffed. Some JDs are AI-generated and contain irrelevant keywords. Don't rewrite your resume to match irrelevant keywords just because they appeared in a bad JD.

The way to avoid these traps: trust the scanner for hard-skills matches, and use your judgment for soft-skills and industry-specific phrases.

What to do after a successful scan

Once your match score is over 75% and you've added keywords honestly:

1. Save the tailored version with a JD-specific filename. "Resume_BackendEngineer_Stripe_2026.pdf"
2. Submit to the actual job posting. ATSes parse on submission, so the version you submit matters; tailored versions score higher than generic.
3. Track the application in a tool like Teal or a spreadsheet, including the match score before and after rewriting.
4. Apply to the next job and repeat. This is most of the active-job-search work.

Closing

Keyword scanning isn't a cheat code; it's tablestakes. Every senior IC and exec I know who's job-hunted in 2026 ran their resume through a scanner against every JD they applied to. The 5-10 minutes per application is among the highest-leverage time you'll spend.

You can scan your resume against any JD for free — no card required to use the diagnostic.

---

Related reading:

• Why Your Resume Gets No Responses
• Complete Guide to Resume Keywords
• How to Tailor Your Resume to Any Job Description
• Best AI Resume Optimizers in 2026