One edit updates the site, the CV and what search engines read
Experience, projects, services and skills are stored once, in English and Bulgarian. The public pages, the PDF CVs, the sitemap and the files for AI assistants are all built from that same data.
The problem
A consultant usually keeps the same facts in four places: a CV in Word, a LinkedIn profile, a website and a text pasted into job portals. After a year they disagree on dates, job titles and technologies, and a recruiter notices.
How dok.sh solves it
Everything is structured data: each position, project, case study, service and skill is a record with an English and a Bulgarian value. Pages only display that data. The same records produce:
- the public site in English and, under
/bg/, in Bulgarian; - PDF CVs in three variants, two designs and two languages;
- the sitemap, page titles, descriptions and JSON-LD for search engines;
llms.txt,llms-full.txtand a Markdown version of each page for AI assistants.

flowchart LR
D[(Resume, projects,<br/>services, skills)] --> W[Website EN + BG]
D --> P[PDF CVs]
D --> S[Sitemap, titles, JSON-LD]
D --> L[llms.txt, Markdown]
The result
Changing a job title or a project description is one edit. The website, the PDF and the search results agree from the next request.
Technologies
Java, Spring Boot, Angular, MongoDB, Content model
- Published: 2026-11-03
- Project: dok.sh — Portfolio and CV Platform
- Case study