Free to read · no account needed

18 min · writing

Rewrite your resume against a JD—without inventing experience

Goal: Mirror a job description’s language while keeping every claim interview-proof.

Best for: Anyone tempted to let AI “improve” a resume by adding skills they do not have.

AI can tighten bullets. It must not invent employers, titles, or metrics. Honesty is the product.

Why beginners skip this

Skipping the hard part of “Rewrite your resume against a JD—without inventing experience” feels fine until a live thread pays the bill. Hold: Mirror JD language; never ship a metric you cannot say aloud—one live pass/fail, or it was only a title Mid-week cost shows up as rework, not as a missing bookmark.

Step 1 — Freeze your evidence bank

Before any rewrite, list only what you can defend: employers, dates, tools you personally used, and metrics you can source (dashboards, emails, shipped links). If a metric is fuzzy, write [METRIC]—do not guess.

Evidence bank (fill offline):
- Roles: [title · company · dates]
- Tools I truly used: [list]
- Proof points: [bullet + where the proof lives]
- Metrics I can show: [number + source] OR [METRIC]

Step 2 — Extract JD phrases (without copying lies)

Ask the model to list skills and verbs from the posting. You will reuse phrasing only where your evidence bank already matches. Matching language ≠ inventing scope.

From this job description, list:
1) Required skills (exact phrases)
2) Repeated verbs
3) Outcomes they care about
Do not rewrite my resume yet.
JD: ---
[paste]
---

Step 3 — Rewrite with an honesty table

Do this step with a real artifact open—not from memory.

Goal: Rewrite my bullets to mirror the JD’s language without adding experience.
JD phrases to prefer: [from step 2]
My bullets: ---
[paste]
---
Format: table with columns:
Old bullet | Rewritten | JD phrase reused | New claims? (must be No) | Interview proof (where)
Boundaries: If my bullets lack evidence, write [NEED EVIDENCE] instead of inventing. No new employers, titles, tools, or metrics.
Example

Filled prompt

JD phrases: “cross-functional updates”, “SQL dashboards”, “experimentation support”
Old: “Made weekly charts for growth; told PM what changed; helped with test sheets”
Rewritten direction: “Delivered weekly SQL dashboards for growth; sent cross-functional updates to PM; supported experimentation with shared test sheets”
New claims?: No — charts→dashboards only if you actually used SQL/Looker; otherwise keep “charts” and reuse other JD phrases you truly match.

What good output looks like

  • Every New claims? cell is No
  • Interview proof column points to a real file or channel
  • Tempting upgrades (“owned roadmap”) are rejected
  • Pass/fail dated (Asia/Shanghai) sits next to the artifact

If the model invents on the resume

Example

Filled prompt

JD wants "led team of 10." My proof: mentored 3 interns.
Mirror language honestly without inventing headcount. Bullets only.

What good output looks like

  • Every metric maps to evidence you own for “Mirror a job description’s language while keeping ”
  • No invented employers, awards, or contacts
  • STAR/result stays sayable aloud in 60s
  • Pass/fail: cut one unsourced claim

Step 4 — Human gate before you apply

Print or split-screen the honesty table. For every Rewritten bullet, say aloud how you would prove it in two sentences. If you stumble, shrink the claim. AI cannot sit the interview for you.

  • Circle any tool name you could not demo tomorrow.
  • Replace empty [METRIC] with a real number or delete the metric clause.
  • Save the table next to the JD PDF—future you will need it when the recruiter calls.
Example

Filled prompt

Gate script (30 seconds per bullet):
“I said ‘delivered weekly SQL dashboards.’ Proof: Looker folder X, every Monday 9:00 to growth Slack. I did not claim I designed the warehouse.”

What good output looks like

  • Proof points to a real place
  • Scope is humble and accurate
  • No borrowed glory from teammates’ work

Refusal script (keep handy)

Stop. Do not invent employers, titles, dates, or metrics.
Rewrite using ONLY bullets I pasted. Mark gaps as [NEED FACT FROM ME].

Red flag

Numbers you cannot source; invented experience—delete what you cannot say in interview. If you cannot explain the refuse line in one breath, stop and rewrite the card before the next run.

Save this card

Card: Mirror JD language; never ship a metric you cannot say aloud—one live pass/fail, or it was only a title. Keep this beside the tool; read before you start.

1. What I shipped: ___ 2. What I refused: ___ 3. What I will never-do again: ___ (date: Asia/Shanghai)

  • [ ] Tools/writes paused
  • [ ] Refuse-line dated
  • [ ] Artifact path saved
  • [ ] Next link chosen (or stop)

Before you start

Open the JD and your evidence bullets side by side for “Mirror a job description’s language while.” Cross out any metric you cannot say aloud in 10 seconds. Guide: career-resume-without-inventing. If you only have a blank practice box, stop and fetch the real job first.

  • Goal in one line: ___
  • Allowed tools: ___
  • Forbidden: ___
  • Human approve needed for: ___

How to use this week

Open one real JD. Mirror its language only onto work you can prove. Ban new metrics from chat. Highlight every number in the draft and source it or delete it. on a live artifact—not a blank demo. Write one refuse-line and date the note (Asia/Shanghai). If you only skimmed career-resume-without-inventing, put 25 minutes on the calendar and redo once.

Pitfalls

PitfallBetter
Letting AI fill empty metric bracketsYou type numbers from real sources only
Keyword stuffing tools you touched onceList tools you could demo live
One resume for every JD with no honesty tableOne table per application; archive it
English JD + Chinese resume without a localize passSee bilingual HK workflow for tone
Example

Filled prompt

Temptation: skip the habit for career-resume-without-inventing to “save five minutes.”
Job still: Mirror a job description’s language while
Do the smallest correct pass: Done visible · one checkable artifact · Unknowns labeled [?].
Boundaries: draft only; no silent send; no invented metrics.
Return: the artifact + the refuse-line you said aloud.

What good output looks like

  • Names the shortcut you refused (specific to career-resume-without-inventing)
  • Still delivers a checkable artifact for “Mirror a job description’s language while”
  • Unknowns labeled; draft-only respected
  • Pass/fail: refuse-line written and dated (career-resume-without-inventing #4)

Practice now

Live card: one bullet rewritten with proof; strike any unverifiable number (Asia/Shanghai).

Today: rewrite five bullets for one JD using only evidence you can say aloud. Paste with “do not invent metrics.” Delete any number you cannot source in ten seconds.

Today card (copy) — career-resume-without-inventing: 1. Artifact I will produce: ___ 2. Pass/fail check: ___ 3. Refuse-line if tempted to skip verify/send: ___ (date: Asia/Shanghai) 4. One never-do from this guide: ___

JD mirror check: highlight three phrases taken from the JD that appear in your bullets—and zero metrics you cannot defend aloud (date: Asia/Shanghai).

  • [ ] Artifact path written
  • [ ] Refuse example would say "yes"
  • [ ] Dated Asia/Shanghai

One sentence takeaway

Mirror JD language; never ship a metric you cannot say aloud—one live pass/fail, or it was only a title

  • You can point to a dated refuse-line
  • You have one artifact path from tonight
  • You can teach the red flag in one sentence

Next

Letter · Slop · Path · Verify before you share · Eval your own outputs

Pair with today: open the first Next link (career-cover-letter-and-mock-interview) only after you dated a refuse-line for this guide (Asia/Shanghai).

← Back to home · Learn · Tools · Compare · Start