How I Build and Publish My CV Automatically

Intro

I keep my CV in a dedicated GitHub repository and publish it in two ways:

1. A webhook server updates the hosted CV file path used by Nginx.
2. My personal site repo pulls the latest CV into public/cv.pdf using GitHub Actions.

The goal is simple: I update CV source once, and the published PDF stays current without manual file copying.

Flow Diagram

               (A) Build Source of Truth
+--------------------------------------------------+
| Repo: giraycoskun/cv                            |
| - Edit LaTeX (cv_data.tex, cv_base.tex, etc.)   |
| - Run: make all                                 |
| - Output: cv_long_YYYY-MM-DD.pdf                |
+-------------------------+------------------------+
                          |
                          v
            +-------------+--------------+
            | GitHub push to cv/main     |
            +-------------+--------------+
                          |
         +----------------+----------------+
         |                                 |
         v                                 v
 (B) Webhook Server Path             (C) Personal Site Sync Path
+-------------------------------+     +------------------------------------+
| webhook-server/scripts/cv.sh  |     | .github/workflows/update-cv.yml    |
| - find latest dated cv_long   |     | - schedule + manual dispatch        |
| - download + validate PDF     |     | - fetch latest dated cv_long        |
| - update symlink              |     | - download to public/cv.pdf         |
|   /media/images/cv.pdf ->     |     | - commit only if changed            |
|   /media/images/cv/pdf/*      |     +------------------+-----------------+
+---------------+---------------+                        |
                |                                        v
                v                           +------------+------------------+
   +------------+-------------------+       | Repo: giraycoskun-com         |
   | Nginx serves stable cv.pdf     |       | Updated public/cv.pdf         |
   +------------+-------------------+       +------------+------------------+
                |                                        |
                v                                        v
       CV available at:                         CV available at:
    https://cv.giraycoskun.dev              https://giraycoskun.com/cv.pdf

Repositories and Components

• CV source repo: giraycoskun/cv
• Webhook script: webhook-server/scripts/cv.sh
• Personal site: giraycoskun-com
• Site sync workflow: .github/workflows/update-cv.yml

I use date-stamped files like cv_long_2026-03-04.pdf and keep a stable public path cv.pdf.

Step 1: Build CV Variants in the CV Repo

The CV repo contains four LaTeX variants:

• cv_short.tex
• cv_long.tex
• cv_short_photo.tex
• cv_long_photo.tex

Build is handled by Makefile:

make all

This generates:

• cv_short_YYYY-MM-DD.pdf
• cv_long_YYYY-MM-DD.pdf
• cv_short_photo_YYYY-MM-DD.pdf
• cv_long_photo_YYYY-MM-DD.pdf

The Makefile also prunes old generated PDFs, which keeps the repo clean and makes “latest file” detection deterministic.

Step 2: Webhook Server Updates Hosted CV

For my self-hosted path, I run a webhook server that executes scripts/cv.sh when my CV repo is updated.

The script does this:

1. Queries GitHub API tree for the CV repo branch.
2. Finds the latest cv_long_YYYY-MM-DD.pdf.
3. Downloads it to a storage folder (if not already present).
4. Verifies it is a PDF (%PDF- header check).
5. Updates a stable symlink used by Nginx.

Core idea from the script:

latest_file="$(
  printf '%s' "${tree_json}" \
    | grep -Eo '"path"[[:space:]]*:[[:space:]]*"cv_long_[0-9]{4}-[0-9]{2}-[0-9]{2}\.pdf"' \
    | sed -E 's/^"path"[[:space:]]*:[[:space:]]*"([^"]*)"$/\1/' \
    | sort \
    | tail -n 1 || true
)"

And then:

ln -sfn "${target_file}" "${SERVE_FILE}"

In my setup:

• CV_STORE_DIR defaults to /media/images/cv/pdf
• SERVE_FILE defaults to /media/images/cv.pdf

Nginx serves the stable file path while the symlink points to the latest dated PDF.

Step 3: GitHub Actions Syncs CV Into My Personal Site

I also keep public/cv.pdf inside my personal site repo so /cv.pdf always works directly from the static site.

The workflow in .github/workflows/update-cv.yml:

• runs daily (cron: "0 6 * * *")
• also supports manual runs (workflow_dispatch)
• fetches latest dated CV filename from giraycoskun/cv
• downloads it to public/cv.pdf
• commits only if file content changed

The fetch step:

latest="$(
  curl -fsSL https://api.github.com/repos/giraycoskun/cv/contents \
    | jq -r '.[].name' \
    | grep -E '^cv_long_[0-9]{4}-[0-9]{2}-[0-9]{2}\.pdf$' \
    | sort \
    | tail -n 1
)"

The update step:

curl -fsSL "https://raw.githubusercontent.com/giraycoskun/cv/main/${latest}" -o public/cv.pdf

Then it commits with github-actions[bot] if git diff detects changes.

Why I Use Both Paths

• Webhook server path is near real-time after a repo update and works well with my self-hosted Nginx setup.
• GitHub Actions path keeps my portfolio repo self-contained and deploy-friendly for static hosting at Cloudflare Pages.

Both use the same source of truth: date-stamped PDF artifacts from the CV repo.

Operational Notes

• The workflow cron is in UTC, so local execution time changes with daylight saving.
• Scheduled GitHub Actions are best effort; manual dispatch is useful when I need immediate sync.
• If the CV repo becomes private, both webhook script and workflow can use GITHUB_TOKEN.
• Keeping a stable public path (cv.pdf) avoids broken links on my portfolio and external references.

Minimal End-to-End Flow

1. Update LaTeX files in giraycoskun/cv.
2. Run make all.
3. Commit and push generated date-stamped PDFs.
4. Webhook server updates hosted symlinked cv.pdf.
5. Site workflow syncs public/cv.pdf and commits the update.

This gives me reproducible CV builds and automated publishing with very little manual work.