AI Resume Maker

Start from scratch or paste an existing resume — and get it built, scored, rewritten, tailored to a specific job posting, and stress-tested against the AI systems that now read resumes first. Eight workflows cover the full journey: build or diagnose, fix every issue, match the target role, simulate the AI screen, write the letters, and export a ready-to-submit file localized to your target market. You bring the career history; it interviews you for the evidence and handles the professional writing, compliance, and formatting.

When to Use / Best For

• You don't have a resume yet: a staged interview (short batched questions, never an interrogation) turns your background into a complete, structured resume — with a dedicated path for new grads and career changers.

• You have a resume but don't know if it's competitive: the score workflow gives you a 100-point diagnostic with honest hard gates — a photo on a US resume caps the grade no matter how good the writing is, because that's what actually happens in the funnel. The score won't flatter you.

• Your resume lists duties instead of achievements: "Responsible for managing a team" is a duty. This skill doesn't just rewrite it — it interviews you for the real numbers (team size, baseline, what changed, who noticed) and rebuilds the bullet from evidence. If you don't have a number, it never invents one.

• You're applying to a specific role: the customize workflow breaks the JD into semantic concept clusters, maps every requirement to evidence in your resume, mines you for matches you didn't realize you had, and tracks coverage as honest concept counts — never a made-up percentage.

• You're applying across borders: one verified fact base → fully localized variants per market. Photo and personal-data rules flip between markets (mandatory in Taiwan-local applications, a liability in the US); Singapore needs a work-pass line; Germany has Lebenslauf and Zeugnisse conventions; Japan uses a two-document set; Korean 자기소개서 pipelines run industrial AI-text detection. 18 markets, each with local vs foreign-MNC tracks, encoded with law/convention/shifting tags.

• You keep applying and hear nothing back: the AI-screen simulation shows you what the machine sees — the parse-layer view of your document, a requirement-by-requirement evidence table with an A–D rating, and a defensibility audit where every weak claim comes with "an AI interviewer might ask you this."

• You need the letters too: market-aware cover letters (including German Anschreiben with DIN 5008 conventions), and reference-letter drafts written for your real referee to review, edit, and sign — including decoding the German Arbeitszeugnis grading code. Drafting is ours; the endorsement always belongs to a real person.

• You need a specific file format: PDF, Word, HTML, LaTeX, Markdown — plus a self-contained, mobile-friendly web resume page you can host as a personal link.

How It Works

Every conversation starts with triage (build from zero / improve / match a JD / test against AI / letters / just convert the format) and positioning (target market × employer type × industry × level). You always see a one-line roadmap of your journey and numbered next steps — no commands to memorize.

Score — 100 points across five dimensions: Content & Evidence (30), Structure & Parser Safety (20), Language & Authenticity (15), Market Fit & Compliance (20), Credibility & Defensibility (15). Letter grade, top strengths with quoted examples, priority-ranked fixes with before/after rewrites, and hard-gate caps for violations that would sink you regardless of prose quality.

Polish — a three-reader checklist: the parser (single column, standard headings, normalized dates), the AI screener (evidence for every claimed skill, coherent narrative), and the human (11-second scan, mobile-first top 20%, zero AI-slop wording). Weak content triggers a mining round before any rewriting. Every change is priority-tagged: critical → major → minor → suggestion.

Customize — JD concept extraction with synonym clusters, a gap matrix with evidence quotes, natural keyword alignment (≤1 JD keyword per sentence — screeners flag stuffing), market-specific must-haves (work-permit lines, salary conventions), plus cover-letter talking points and interview prep notes for remaining gaps.

AI-Screen — simulates a 2026-style screening pass: what the parser extracts (and what your layout loses), a rubric-based evidence check with an A–D rating, and a line-by-line defensibility audit. Clearly labeled as a diagnostic simulation — real ATS never show candidates a score, and this skill won't pretend otherwise.

Export — a selection tree picks the right template from market × industry × level:

Paper size (Letter/A4), photo handling, date formats, spelling variants, and file naming all follow the target market. Every delivery passes a pre-flight check (parseable text layer, working links, page caps, no placeholder residue) — reported honestly, including anything the environment couldn't verify.

Why This Skill / What Makes It Different

Excavation, not fabrication: most AI resume tools hallucinate impressive numbers. This one refuses — including "just put a placeholder," "it's only a test," and "my friend said it works." Instead it runs a structured interview (six question dimensions, ten role-specific packs, narrative packs for 자기소개서/自傳) that surfaces real competitiveness you didn't think to mention. Anything unverified ships as a visible [needs evidence] marker, never as a confident fake.

Built for AI readers and human readers: parser-safe structure, semantic JD coverage for LLM screeners, and an 11-second human scan test — with reader priority weighted per market (machine-first US pipelines, platform-first 104/Boss直聘 ecosystems, human-first German Mittelstand).

No fake scores: coverage is counted, not invented; the AI-screen rating is evidence-backed and labeled as a simulation; the diagnostic score carries hard gates and a disclaimer. If it can't verify something, it says so.

Market rules as law, not trivia: 18 markets with legal citations where they exist (EEOC, Equality Act, TAFEP, 채용절차법, 就業服務法), local vs foreign-MNC dual tracks, and per-market naming conventions down to the delivery filename.

Transparent audit trail: every checklist item is marked, every change is priority-ranked, every mined number traces to your own answer.

Languages

Coach in English or Chinese (Traditional and Simplified both mirrored — Taiwan users get Traditional with local vocabulary). Documents are written in the target market's expected language, including bilingual sets where the market expects both (Taiwan tech: zh + en).