Masterful CV
A live CV and application-workflow product for turning existing career information into a structured, editable document that can be checked before print or PDF export.
The job to be done
People often arrive with an existing CV, not a clean data model. Masterful CV turns that material into structured profile information while keeping the document visible, editable, and exportable. The aim is a precise document workflow rather than a prompt box that silently rewrites career history.
Try the live workflow
The strongest proof is the live product. A useful inspection takes only a few minutes:
- Import an existing CV or start from a blank profile. You can import PDF, DOCX, text, HTML, Markdown, or profile JSON.
- Edit a structured section and confirm that the document preview changes with it.
- Preview the document live.
- Export for print or PDF.
That path exercises the product's main claim: profile data is the canonical source for editing, live preview, and document generation. The preview is not a disconnected mock-up of the output.
What I own
I design and build the product end to end: discovery, information architecture, document and application workflows, interface implementation, APIs and data model, identity, billing, AI-assisted import, localization, support surfaces, testing, and release work.
Decisions, alternatives, and costs
Keep one document model
Decision: profile data remains the source for the editor, preview, print route, and import flow.
Alternative: let each surface keep its own convenient copy and reconcile them later.
Cost: the shared model needs careful normalization, validation, migration, and error handling. That work is less visible than a new screen, but it prevents the preview and exported document from quietly disagreeing.
Make AI import propose, not replace
Decision: imported material becomes structured review material before it can update the document.
Alternative: accept the model response immediately and optimize for a shorter path.
Cost: review adds a step, and partial or warning-backed results need a clear interface. I accept that friction because a plausible-looking rewrite can still omit, invent, or misclassify important career information.
Allow useful work before sign-up
Decision: a visitor can create and edit an anonymous draft in browser-local storage.
Alternative: require an account before the editor becomes useful.
Cost: a local draft cannot follow the user across devices, and the boundary between local and cloud persistence has to be explained clearly. The benefit is that the core workflow can be evaluated before identity or billing gets in the way.
Product scope
The document loop sits inside a broader product: identity and onboarding, saved profiles, job and application tracking, billing and credit views, account and legal flows, admin support surfaces, localization, and automated browser, unit, and accessibility checks. Those parts support the document rather than replacing it as the product's centre of gravity.
Data and AI boundaries
Anonymous drafts are stored locally in your browser and are not cloud-synced. AI-assisted import is a separate flow and may process imported material server-side. Its structured result, validation warnings, and proposed content are shown before they change the working document; direct identifiers are redacted before document content is sent to an AI provider.
Signing in adds saved profiles, jobs and applications, cloud data, paid plans, and credits around the same document loop.