PAIN 01
One resume, many roles
Tailoring for every job is slow, manual work, so most people send one generic version everywhere.
Vantage AI · Career co-pilot · Self-initiated
One resume. Every role. Told your way.
An AI career co-pilot that turns one resume into a role-specific story: tailored to each job, styled to who you are, and approved by you line by line.
Sole product designer and builder
Research to working build · Claude Design, Claude API, Streamlit, Python
ROLE
Sole product designer and builder
TOOLS
Claude Design · Claude API · Streamlit · Python
TYPE
Self-initiated and solo, no client brief
FOR
Multi-role applicants, career switchers
01 · Problem statement
Job seekers rewrite the same resume dozens of times, and still can’t tell why it isn’t landing.
PAIN 02
Scores without guidance
Keyword scanners return a number and missing terms, not how to tell the story so a recruiter sees the fit.
PAIN 03
AI that flattens voice
Chat assistants produce polished but interchangeable bullets, overwrite your voice, and can overstate what you did.
Problem statement
Job seekers applying across multiple roles need a fast way to tailor each resume, see exactly where they fall short, and keep their own voice, because today’s tools either score without guiding or generate without understanding who the person is.
WHO IT IS FOR
Multi-role applicantsCareer switchersInternational candidatesPeople with non-linear paths02 · How might we
Five questions that shaped every screen
HMW 01SPEED
How might we tailor a resume to each role in minutes rather than hours?
HMW 02CLARITY
How might we turn a match score into specific, actionable next steps?
HMW 03CONTROL
How might we let AI do the heavy lifting while the person stays the final decision-maker on every line?
HMW 04IDENTITY
How might we reflect who someone is, not just which keywords they have, in how their resume looks and reads?
HMW 05CONTINUITY
How might we keep tailoring, exporting and tracking in one connected flow?
03 · Research · comparative analysis
Existing tools solve one slice. None connect the whole job.
| ATS keyword scanners | AI resume builders | General chat assistants | Vantage AI | |
|---|---|---|---|---|
| Match score against a specific JD | Core | Varies | Manual | Core |
| Explains gaps with next steps | Partial | Partial | If prompted | Core |
| Rewrites bullets | Rare | Core | Core | Core |
| You approve every change | n/a | Rare | Manual | Core |
| Style reflects the person | No | Templates | No | Archetypes |
| Application tracking | Varies | Varies | No | Core |
| ATS-safe export | Core | Varies | No | Core |
INSIGHT 01
Scanners diagnose but do not treat: the user still has to work out how to fix the resume.
INSIGHT 02
Builders and chat assistants treat but do not diagnose, and rarely ask the user to approve each change.
INSIGHT 03
Nobody starts from who the person is. Every tool starts from a template or a blank prompt.
04 · Ideation
From a resume fixer to a career co-pilot
Ideas grouped into four pillars; anything that did not serve one was cut.
PILLAR 01
Identity
A three-question Identity Quest maps each person to an archetype that sets their resume’s type, palette and layout.
PILLAR 02
Intelligence
The Forge reads the job description, scores role fit, surfaces skill gaps and rewrites bullets with the X-Y-Z formula.
PILLAR 03
Control
Nothing is applied silently. Every suggestion can be accepted, edited or rejected.
PILLAR 04
Continuity
The Vault exports PDF, DOCX and ATS-safe TXT plus a live link, and hands off to the Job Tracker.
CORE FLOW
PortalIdentity QuestRevealThe ForgePreviewVaultJob Tracker
FOUR ARCHETYPES
Architect
Grid, Swiss, data-heavy
Visionary
Editorial, serif, bold
Specialist
Monospace, technical
Humanist
Warm, rounded, story-first
05 · Why this works · journey mapping
The same journey, without the dips
Proto-persona: the multi-role job seeker. Dashed = today, solid = with Vantage.
01
Find a role
02
Tailor resume
03
Check the fit
04
Apply
05
Track and follow up
DOING
Saves roles across job boards
Rewrites bullets by hand for each JD
Runs a keyword scan, guesses at fixes
Exports, reformats for each portal
Keeps a spreadsheet, loses track of versions
THINKING
“Which of these am I actually a fit for?”
“This takes hours per application.”
“The score went up but I don’t know why.”
“The portal mangled my formatting.”
“Which resume did I send them?”
WITH VANTAGE
Role-fit score before you commit time
The Forge drafts from the JD in one pass
Gaps explained, X-Y-Z rewrites to approve
PDF, DOCX and ATS-safe TXT from the Vault
Tracker keeps the resume version per application
06 · UX wireframes
Lo-fi, mid-fi, hi-fi. Seven screens, three passes.
Every screen was drawn three times, from structure to a dark, finished design system.
Portal
LO-FI
MID-FI
HI-FI
Reveal: your archetype
Drop image
va-lofi-03-reveal.png
va-lofi-03-reveal.png
LO-FI
MID-FI
HI-FI
Preview: score, gaps, approve each line
LO-FI
MID-FI
HI-FI
07 · Hi-fi
The full co-pilot, from Portal to Job Tracker
A dark design system in Fraunces and IBM Plex, built as a working app.
08 · Prototyping with Claude
Claude as a design partner, not an autopilot
Claude Design to explore and prototype the experience; the Claude API as the engine behind The Forge.
01
Frame
Mapped the six-step flow and each screen’s job before prompting, so the AI worked to a structure.
02
Prototype
Prototyped the multi-screen co-pilot in Claude Design, with one coherent design system across screens.
03
Prompt
Designed the coaching prompts. Framing the model as an expert career coach gave far more specific feedback.
04
Build
Built the working app on the Claude API and Streamlit, then redesigned it with a dark design system.
WHAT WORKED
A consistent visual system across every generated screen, and a fast loop from idea to clickable flow.
MAIN CHALLENGE
Scope creep. Every iteration invited another feature. I locked the six-step core flow and parked everything that did not serve the four pillars.
09 · What problem it solves
From guesswork to a guided, owned process
BEFOREWITH VANTAGE AI
Hours of rewriting for each application
One pass from job description to tailored draft in The Forge
A score with no explanation
Role-fit score plus named gaps and specific bullet rewrites
AI that overwrites your voice
Every change is proposed; you accept, edit or keep your own
A resume that looks like everyone else’s
An archetype-driven layout that reflects how you work
Versions scattered across folders and spreadsheets
Vault exports and a tracker that remembers what you sent
10 · How it is better
Other tools optimise a document. Vantage coaches a person.
01
Starts with identity, not a template
The Identity Quest sets how the resume looks and reads before a bullet is written.
02
Human in the loop by design
AI guides and analyses; the person makes every call. Nothing is applied silently.
03
Explains, not just scores
Every score comes with the gaps behind it and a concrete way to close them.
04
JD-aware, measurable rewrites
X-Y-Z bullets against the specific role, with placeholders instead of invented numbers.
05
End to end in one place
Tailor, export PDF, DOCX or ATS-safe TXT, share a live link and track the application.
Designed, prototyped and built end to end by me as a self-initiated project.
Neha Mayacharya · Product & Experience Designer
MORE PROJECTS
All work
