Industrialized AI Token Platform — Unified Journey

Single learning journey that combines nine enablement modules into one course for deploying an industrialized AI token generation platform on Red Hat OpenShift AI 3.x. Focus: optimized resources, performance SLOs, model governance, and engineering outcomes.

This repository was created from the course-starter-template. For first-time setup (clone, course-init, GitHub Pages), see README-TRAINING.md or the template link above.

---

Course content: the nine modules

Repositories pulled in (content stays in those repos; this repo provides the playbook and journey map):

• Phase 1: rhoai3-storage, rhoai3-registry, rhoai3-catalog
• Phase 2: rhoai3-hwprofiles, rhoai3-gpu-aas
• Phase 3: rhoai3-deploy
• Phase 4a: llmops-llmd (distributed inference — required for MaaS)
• Phase 4b: rhoai3-maas
• Phase 5: rhoai3-validate

---

Prerequisites for building

• All nine content repos must be cloned as siblings of this repo (same parent directory). Example layout:

  parent/
  ├── rhoai3-journey/     ← you are here
  ├── rhoai3-storage/
  ├── rhoai3-registry/
  ├── rhoai3-catalog/
  ├── rhoai3-hwprofiles/
  ├── rhoai3-gpu-aas/
  ├── rhoai3-deploy/
  ├── llmops-llmd/
  ├── rhoai3-maas/
  └── rhoai3-validate/
  

• Node.js (for npx antora) or Docker (for antora/antora image).

---

Build the site

Option 1: Docker (recommended)

Run from the parent directory that contains rhoai3-journey and all nine content repos:

cd /path/to/parent
docker run -u $(id -u) -v $PWD:/workspace:Z --rm -t -w /workspace/rhoai3-journey antora/antora antora-playbook.yml

Then open rhoai3-journey/build/site/index.html.

Option 2: Local NPM

cd rhoai3-journey
npm install
npx antora antora-playbook.yml

Then open build/site/index.html. Run from inside rhoai3-journey with the other repos as siblings.

---

Why the nav links work (or don't)

The playbook uses git URLs for the nine content repos so that the GitHub Pages build (and any CI build) fetches and builds all of them. That way the generated site includes every component (e.g. rhoai3-registry/1/chapter1/section3.html), and the journey nav xrefs resolve to real pages instead of fragment fallbacks. If the playbook only had url: ./, only the journey pages would be built and nav links would change the URL to a #... fragment and not load the target page.

For local builds with sibling clones, you can temporarily point sources at local paths (e.g. url: ../rhoai3-storage) in the playbook to avoid cloning; the same playbook with git URLs also works locally (Antora will clone the repos into its cache).

---

Repository structure (journey-specific)

• antora-playbook.yml — Build config; lists this repo plus nine content sources; uses template UI (./ui-bundle, ./supplemental-ui).
• antora.yml — Journey component descriptor (ROOT module only for the journey map).
• modules/ROOT/ — Journey Map (index.adoc) and Phase 1–5 pages (phase1.adoc … phase5.adoc, phase4a.adoc, phase4b.adoc).
• TASKS.MD, INSIGHTS.MD — Task log and insights for the unified journey.

Template files (e.g. LABENV, chapter1–3, appendix, images, .github/workflows) remain from the starter template; the built site is driven by the playbook and ROOT content above.

---

Build notes

• Start page: The site is configured to open at the Journey Map (rhoai3-journey::index.adoc). If it is not found, open build/site/rhoai3-journey/1/index.html after a successful build.
• Errors in other repos: AsciiDoc errors (missing images, xrefs, doctype) in the upstream rhoai3-* and llmops-llmd repos must be fixed in those repositories.

---

See also

• README-TRAINING.md — Template and training repo setup.
• Development using devspace
• Guideline for editing content