resume.jnap.me — editorial dossier
resume.jnap.me — editorial dossier
Astro single-page resume that reframes the resume-site/ content into an editorial dossier for creative firms. Typography leans on Fraunces + Space Grotesk, with a controlled grid and two intentional "weird" interruptions (rotated systems note + section code overlays) to nod to Wolff Olins' Lloyd's project without copying it.
Stack
- Astro 5 static output (Vercel friendly)
- TypeScript + module aliasing for cleaner imports
- PDF rendered from the canonical public web resume
- Single CSS system in
src/styles/global.csswith embedded print + motion rules - Systems section has layout + view toggles (grid vs. column stack for highlights; accordion/table/badges/tabs/timeline for the catalog).
Getting started
npm install # or make install
npm run dev # start Astro locally on http://localhost:4321
npm run build # production build into ./dist
npm run preview # serve the build (used for PDF export)
npm run skill:graph -- --clusters=6 # generate skill clusters via OpenAI embeddings
make wrappers are included for parity (make dev, make build, make pdf, etc.).
Content + configuration
- All resume facts live in
src/data/resume.json(copied directly from the providedresume-site/). Do not edit them inline in the markup. - Hero/section copy lives in
src/data/copy.jsonso the prose can be tuned without touching templates. - CTA + site metadata live in
src/config.ts. UpdatebookCallUrlthere if the scheduling link changes. - The hero "Download PDF" button serves
public/resume.pdf. Regenerate it whenever the site changes (see below) so Vercel ships the current print.
Generating the PDF
make pdf
This runs npm run pdf: build the canonical JSON-backed page, then print its actual public DOM through scripts/export-pdf.mjs to public/resume.pdf. This preserves the web contact privacy rules and current career facts. npm run pdf:html exports an already-built page. Do not replace this artifact with the separate LaTeX/pit resume: its facts and phone policy differ. The old fetch script is retained for historical reference and is not the default export path.
Deploying to Vercel
- Push to GitHub and import the repo in Vercel.
- Framework preset: Astro.
- Build command:
npm run build; Output directory:dist. - Ensure
public/resume.pdfis committed so the CTA stays live.
Typographic + layout system
- Fraunces headlines use optical-size variations layered over Space Grotesk body copy to echo editorial dossiers.
- 12-column-inspired grid with alternating spans for Experience vs. detail rails.
- Section codes offset in the margin and the rotated hero systems-note are the two intentional "weird" moves.
- Skill meters visualize years of experience (normalized across the full catalog) instead of raw numerals only.
- Print styles strip shadows/backgrounds, remove CTA buttons, reveal URLs, add a masthead, and include page numbers.
Verification checklist
- ✅ Book-a-call CTA hits
https://cal.com/justin-napolitano-gvu3p3/intro. - ✅ Download PDF points to
/resume.pdf(regenerate viamake pdf). - ✅ Experience, projects, skills, education, and publication match
resume-site/content. - ✅ Layout responds across mobile/desktop + print.
Skill graph pipeline
Set OPENAI_API_KEY (the script calls OpenAI's embeddings API) and run:
npm run skill:graph -- --clusters=6
The script reads src/data/resume.json, filters to the displayed skills, generates embeddings via text-embedding-3-small, clusters them with a lightweight k-means implementation, and writes public/skill-graph.json with cluster labels, membership, and 2D PCA coordinates. Pass --hidden=true to include suppressed skills, --model=text-embedding-3-large for higher fidelity vectors, or --output=some/path.json to change the destination. After generating the file, visit /skill-graph (static map) or /skill-graph-test (interactive canvas playground) in the local dev server.