CV Publishing System

Markdown-as-source-of-truth CV pipeline. Edit cv/cv.md, run one command, get HTML + PDF + DOCX.

Quick start

Everything runs in Docker -- no local Pandoc or LaTeX install needed.

# Build all outputs
make build

# With contact info for PDF/DOCX
CV_EMAIL="you@example.com" CV_PHONE="+44..." make build

# Preview locally
make serve
# Open http://localhost:8000

Outputs

Output	Path	Purpose
HTML	dist/index.html	Public page with dark/light theme
PDF	dist/cv.pdf	Recruiter download (Inter font)
DOCX	dist/cv.docx	Editable export

The HTML page includes download buttons and a theme toggle. Email and phone are excluded from the public HTML page.

Project structure

cv/
  cv.md              # Source of truth
  build.py           # Build script
  CNAME              # Custom domain for GitHub Pages
  templates/
    cv.html.j2       # Tailwind HTML layout
    cv.tex           # LaTeX template (Inter, titlesec)
  styles/
    cv.css
Dockerfile           # Pandoc + TeX Live + Inter font
Makefile             # Docker-based build targets
.github/workflows/
  build-cv.yml       # CI: lint, build, deploy to GitHub Pages
.pre-commit-config.yaml  # ruff, uv-lock, rust fmt/clippy
dist/                # Generated outputs (gitignored)

Make targets

make build        # Generate all outputs (HTML, PDF, DOCX)
make html         # HTML only
make pdf          # PDF only
make docx         # DOCX only
make serve        # Build and serve on port 8000
make clean        # Remove dist/
make docker-build # Build the Docker image only

CI/CD

GitHub Actions pipeline on push to main:

1. Lint -- pre-commit checks (ruff, trailing whitespace, yaml, etc.)
2. Build -- Docker build, generate artifacts with secrets injected
3. Deploy -- publish to GitHub Pages at lucaromagnoli.dev

Contact info (CV_EMAIL, CV_PHONE) is stored as GitHub secrets and injected at build time.

Spec

Full specification: SPEC.md.