CV as Single Source of Truth — how a résumé became an automation project

A CV sounds like a simple document. In practice, it's a maintenance problem.

Every change — a new project, an updated technology, a new time period — requires edits in multiple places: the German PDF version, the English one, the personal website, application documents in various formats.

That's four locations for the same information. Four places where inconsistencies arise. The core issue: no single source of truth.

A single data file contains all CV content bilingually. From there, an automated pipeline generates every output — with tech experience years calculated automatically from real project periods.

One change → one command → all outputs up to date and consistent.

The TypeScript logic was written by AI. I defined the structure, AI implemented, I reviewed. No loss of control — a different relationship with implementation.

What used to take days: one afternoon.

This pattern occurs daily in companies: price lists in Excel and on the website, product data in ERP and shop, contract texts in Word and CMS.

The problem is not a CV-specific edge case — it's an architecture problem.