Skip to main content

AI Docs-as-Code Workflow

Governed Docs-as-Code architecture

Type: Public GitHub portfolio project
Repository: AI-Docs-as-Code-MKDocs
Published documentation: Live MkDocs site

Challenge​

Traditional documentation often separates writing, review, publishing, and downloadable outputs into different manual workflows. This project demonstrates how those activities can be governed through the same version-controlled pipeline.

Architecture​

  1. Authors maintain documentation in Markdown.
  2. Changes are managed in GitHub through version control and pull requests.
  3. Vale performs deterministic style and terminology checks.
  4. An AI documentation review adds contextual, consistency, security, and risk findings.
  5. Human review and approval remain part of the publishing gate.
  6. MkDocs Material builds the documentation site.
  7. GitHub Actions publishes HTML and produces PDF and DOCX outputs.

My contribution​

  • Designed the end-to-end workflow
  • Configured Markdown/MkDocs publishing
  • Implemented GitHub Actions automation
  • Integrated Vale style review
  • Added AI-assisted review controls
  • Kept AI findings non-authoritative and human-verifiable
  • Implemented multi-format publishing

What it demonstrates​

The project shows documentation as a software-like, traceable engineering workflow: source-controlled, reviewable, testable, publishable, and reusable.

Open the GitHub repository · Back to flagship projects