Content Authoring

Content authoring should keep one rule in mind: every piece of content should know how it participates in learning.

A post can exist as source material, but platform content should connect to lessons, roadmaps, skill paths, and flashcards.

Authoring Units

The current platform uses these content units:

• posts for original long-form material
• lessons for structured learning pages
• roadmaps for broad learning journeys
• courses for ordered lesson sequences
• skill-paths for narrower goals
• challenges for flashcard decks and review tasks

When adding new ML content, avoid adding only a blog post. Add enough metadata so the content can appear in the learning system.

Minimum Useful Metadata

A lesson should define:

• title
• description
• domain
• track
• module
• order
• difficulty
• estimated minutes
• tags
• prerequisites
• related lessons
• challenge references

This metadata is what allows the site to build dashboards, roadmaps, review queues, and navigation.

Turning an Article Into a Lesson

1. Preserve the original article. Keep the source post route while the new lesson route is being validated.
2. Assign the lesson to a track. Decide whether it belongs to Transformer Systems, AI Agent Frameworks, or a future ML/deep learning track.
3. Extract recall targets. Identify the definitions, mechanisms, comparisons, and traces that should become cards.
4. Place it on a roadmap. Add the lesson to the right stage and define any prerequisites.

Adding New Tracks Later

Future tracks can reuse the same structure.

Examples:

• Deep Learning Fundamentals
• Training Systems
• Evaluation and Alignment
• RAG Systems
• Agentic Application Engineering

Each new track should have its own roadmap and flashcard review layer. That way, the site stays a learning platform rather than becoming another flat content archive.