How to Use Your CV with an LLM: Export It to YAML First

Why YAML beats pasting your PDF into ChatGPT

When you copy a CV out of a PDF and paste it into a chat, you don't paste a clean document — you paste whatever text-extraction order the PDF happens to have. Two-column layouts interleave. Headers and footers land mid-sentence. Dates detach from the roles they belong to. The model then spends effort reconstructing the structure before it can help you, and it sometimes reconstructs it wrong.

A structured format removes that whole problem. The model reads explicit fields — name, experience, each role's title, company, dates, and bullets — instead of inferring them from spacing. YAML is the cleanest structured format for this job, and it's also the most economical: less of your prompt budget is spent on the CV, leaving more room for the actual task.

YAML vs JSON vs plain text for LLMs

All three work, but they're not equal for this use case:

• Plain text (pasted from a PDF) — smallest effort, worst result. The model has to guess the structure, and column-based layouts scramble badly.
• JSON — fully structured and unambiguous, but token-heavy: every key is quoted, every level adds braces and commas. It's also harder for a human to skim and edit before sending.
• YAML — structured like JSON but with far less punctuation. Indentation replaces braces, keys don't need quotes, and multi-line text (your summary, your bullets) reads naturally. It's the sweet spot: machine-clean and human-readable.

How to get your CV as YAML

You don't need to write YAML by hand or know the syntax. If your CV is already built on TakeMeUp.cv, the CV-to-YAML add-on exports a clean, ready-to-paste YAML file in one click — free, no subscription. It drops the layout fields an LLM doesn't need and keeps the structure an LLM does.

If your CV lives elsewhere, you can still build it on TakeMeUp.cv for free (import from LinkedIn, Word, or PDF) and then export the YAML — no payment needed for the export itself.

Five prompts that work best with a YAML CV

Paste the YAML as your first message, then try any of these. Replace the bracketed parts with your own details:

1. Tailor to a job: “Here is my CV in YAML. Rewrite the experience bullets to target this job description, using only achievements already present. Don't invent metrics.” Then paste the job post.
2. Draft a cover letter: “Write a cover letter for [role] at [company] based only on this CV. Keep it under 250 words and specific.”
3. Mock interview: “You're the hiring manager for [role]. Ask me five interview questions based on this CV, one at a time, and critique my answers.”
4. Skill-gap analysis: “Compare this CV against a typical [target role] and list the three skills I'm most missing, ranked by how often they appear in job ads.”
5. Summary rewrite: “Rewrite my professional summary in three different tones — concise, confident, and warm — using only what's in this CV.”

Keep it accurate and private

Two rules. First, you are the fact-checker: LLMs are excellent rewriters but will sometimes round 'led a project' up to 'led a 12-person team' if you let them. Always tell the model not to invent numbers, and read every rewrite against your real CV before you send it to an employer.

Second, mind your data: a CV pasted into a public chatbot may be retained or used for training depending on the provider's settings. Remove your home address and any ID numbers before pasting, use the provider's data-controls / no-training options where available, and treat the chat as you would any third-party service that sees your personal data.