Software engineer postings tend to repeat the language, the place the service runs, and the review habit: Go or another language, Docker, Kubernetes, a database such as PostgreSQL, Redis, GitHub, pull requests, tests, and CI/CD. A preferred line may add TypeScript or AWS without making them the same as the required stack. Include a name only when you built, ran, or reviewed with it.
The posting and the snippet below are fictional. The lists are the current output of the job description keyword extractor and the resume keyword matcher. The score is literal keyword overlap, not an ATS pass rate. Rounding is explained in what an online ATS score actually measures.
Words that recur for this role
Common technical vocabulary, not one stack required everywhere. Name the service, the store, or the review.
| Category | Phrase | How a bullet usually carries it |
|---|---|---|
| Tools or systems | Go | Name the service. The tool records a standalone Go when the posting also names another software tool. |
| Tools or systems | Docker and Kubernetes | Name what you packaged or where it ran. One does not stand in for the other. |
| Tools or systems | PostgreSQL and Redis | Name what you stored or cached. |
| Tools or systems | GitHub | Name the review. The combined list keeps the spelling GitHub. |
| Responsibilities | Pull requests | Name what you reviewed. |
| Responsibilities | REST API | Name the interface you exposed. |
| Responsibilities | Unit testing and CI/CD | Name the check you added or the pipeline you used. |
| Tools or systems | Linux, TypeScript, and AWS | Name the environment. Leave a cloud off when you have not used it. |
A fictional posting in the extractor
Fictional job description. This is not a real vacancy.
Software Engineer Build a backend service in Go and expose a REST API. Package the application with Docker and run it on Kubernetes. Store records in PostgreSQL and cache sessions in Redis. Review pull requests on GitHub and add unit testing for the new paths. Required: CI/CD experience and Linux. Preferred: TypeScript for a small internal page, and AWS.
Top skills stops at 12 and leaves Go off that panel: Pull requests, Unit testing, REST API, CI/CD, GitHub, AWS, Kubernetes, PostgreSQL, TypeScript, Docker, Redis, and Linux. Go is still a real match. Tools and software lists it with the others: GitHub, AWS, Kubernetes, PostgreSQL, TypeScript, Docker, Redis, Linux, and Go. Soft skills find no clear matches. Action verbs: Build. The required panel and the preferred panel each keep one sentence.
The combined keyword list has 13 lines: Pull requests, Unit testing, REST API, CI/CD, GitHub, AWS, Kubernetes, PostgreSQL, TypeScript, Docker, Redis, Linux, and Go. The qualification sentences stay on their panels. Spellings stay GitHub, PostgreSQL, TypeScript, and AWS. The matcher scores 13 phrases and does not score those two sentences as keywords.
The same posting in the matcher
Fictional resume snippet. Not a real engineer.
Software Engineer • Built a backend service in Go and reviewed pull requests on GitHub. • Ran the service with Docker and stored records in PostgreSQL. • Skills: Go, Docker, GitHub, PostgreSQL.
Keyword overlap is 38. The note reads: 5 of 13 job-description keywords also appear in your resume. The matcher rounds 5 ÷ 13 × 100 to 38.
Matched: Go, Pull requests, PostgreSQL, Docker, and GitHub. Important skills found are those same five.
Missing: Unit testing, Kubernetes, TypeScript, REST API, CI/CD, Linux, Redis, and AWS. Suggested keywords are those eight. The snippet never says REST API, so that phrase stays missing. The matcher does not infer it from “backend service.” Leave the eight off, including preferred tools you have not used.
How to match a resume to a job description uses a marketing example for the same limit: a related tool does not satisfy a different word.
A before-and-after bullet
Fictional notes. Built a backend service in Go. Reviewed pull requests on GitHub. Ran it with Docker. Stored records in PostgreSQL. Did not deploy on Kubernetes and did not use AWS.
Before, fictional. “Worked on the backend.”
After, fictional. “Built a backend service in Go and reviewed pull requests on GitHub.” The language and the review tool are in the notes. REST API, Kubernetes, and AWS stay off. No latency figure was invented.
Check a draft you actually wrote
Read the posting with the keyword extractor, then compare the full posting with your resume in the keyword matcher. The resume skills extractor lists languages and tools the resume already names. The tools run in the browser. Pasted text is not sent to a server or an AI service.
Frequently asked questions
Short answers.
If I used Docker, does the resume also match Kubernetes?
No. On this fictional pair Docker matches and Kubernetes stays missing. Write Kubernetes only when you ran or deployed with it.
Does a high overlap score mean I will get an interview?
No. A high score does not guarantee interviews. This fictional pair scores 38, from 5 of 13 phrases. That is literal keyword overlap, not an ATS pass rate and not a hiring forecast.
Is the text I paste sent to a server or an AI service?
No. The keyword extractor, keyword matcher, and skills extractor process pasted text in your browser. It is not sent to a server or an AI service. The privacy policy says the same about the free tools.
The top-skills panel left Go off. Did the tool miss it?
The panel lists at most 12 phrases, and Go was past that cut. Tools and software still lists Go, the combined keyword list includes Go, and the matcher matches it because the resume says Go. Read those panels together before you assume a language was ignored.