Tailor Your Resume Offline with a Local Model

Why this works

Tailoring a resume to each posting is the most effective and most tedious part of a job search. You can do it privately: keep your resume and the job posting on your disk, and let a local model suggest edits you approve line by line.

Step by step

1. Save both documents as plain textCopy your resume and the job advert into resume.txt and job.txt. Plain text is easier for a small model than PDF.
2. Start a local modelRun ollama run llama3.2 and leave the chat open. Nothing you type is sent to a server.
3. Ask for matches, not rewritesUse the prompt below. You want a list of which of your existing bullets map to the job's requirements, and which requirements have no evidence.
4. Write the edits yourselfChoose the bullets to keep, reword them in your own voice, and add numbers only if you can back them up.
5. Run a final consistency checkAsk the model to flag dates, job titles, or claims that differ between the old and new versions.

Copy-paste prompt

Compare my resume and the job description below. Output three lists: (1) requirements my resume already evidences, quoting my bullet; (2) requirements with weak or no evidence; (3) my bullets that are not relevant to this role. Do not invent experience. Do not add numbers that are not in my resume.

RESUME:
<paste>

JOB DESCRIPTION:
<paste>

Check your result

• No claim appears that is absent from your original resume
• Dates and employer names match your source document
• You can defend every bullet in an interview

Pitfalls to avoid

• Local models sometimes invent metrics. If a number is not in your original, delete it.
• Pair this with the ATS and impact-statement guides in this directory for formatting.