AI Knowledge Graph

AI Knowledge Graph is a personal learning system for studying AI engineering through a structured, evolving curriculum. It turns Markdown study notes and a 228-entry syllabus into a browsable knowledge map, searchable chapter library, and in-page note reader.

Its main deliverable is a static browser application backed by generated JSON and HTML. The curriculum covers foundations, APIs, retrieval, agents, multi-agent systems, memory, evaluation, safety, production, and frontier topics.

Explore

• Technical and product blueprint — architecture, data flow, authoring tools, setup, caveats, and source map.
• Curriculum source · Markdown notes · Static application
• Generated graph index · Build pipeline

What is implemented

• A D3-powered graph with concentric curriculum rings and a force-directed view, zoom, pan, labels, search, and node inspection.
• A curriculum library with status, domain, type, ring, and size filters, plus sorting and a “continue” chapter affordance.
• Static note pages and an in-app reader with cross-note navigation, backlinks, code presentation, and deep links.
• A Node.js compiler that parses the curriculum and note frontmatter, resolves [[wikilinks]], builds node/edge metadata, emits JSON indexes, and generates note pages.
• A note validator with curriculum-to-note checks, metadata checks, privacy-pattern scanning, and a minimum-content check for notes marked lit.

The checked-in index reports 228 curriculum entries, 1,522 edges, zero entries marked lit, and all 228 marked charging. These are snapshot values in the committed generated data, not a claim that every planned exercise has been completed. Read the implementation and maturity notes before interpreting the status UI.

Run locally

Requires Node.js. The project declares no npm dependencies. From the repository root:

npm run validate-notes
npm run build-brain
npm run serve

Then open http://localhost:8080. The serve script uses npx http-server, which may fetch the utility if it is not already cached. The app loads D3.js and Marked from CDNs, so a network connection is needed for those libraries. See setup and troubleshooting.

Maturity

This is a substantial curriculum and a working static-site implementation, with supporting authoring and indexing scripts. It is best understood as a personal learning tool and portfolio prototype rather than a production learning platform: progress in the browser is derived from generated note metadata, there is no account or database-backed sync, and the curriculum snapshot currently has no lit nodes. Some adjacent brain/ utilities describe a broader personal assistant and integrations, but they are separate from the static curriculum UI.

Technology

Area	Implementation
Browser UI	HTML, CSS, vanilla JavaScript
Graph and simulation	D3.js v7, loaded from CDN
Note compilation	Custom Node.js scripts and Markdown transforms
Content and syllabus	Markdown with YAML-like frontmatter; curriculum source parsed by project code
Generated artifacts	Static HTML and JSON under site/
Hosting configuration	Vercel output directory set to site/

See PROJECT_BLUEPRINT.md for the data model, algorithms, security boundaries, and verified limitations.

---

This README describes the checked-in source and generated snapshot. It does not claim that a live deployment was verified.