A free, adaptive AI curriculum that meets you where you are. Take a 6-question assessment and get a study schedule tailored to your experience level — from first encounter to frontier research.
Most AI courses skip it: Your AI Data Footprint — what gets stored, what trains the model, and what stays local. It's a privacy sidecar you can run next to any primary track, built from the real data-use disclosures of the tools you already use (Anthropic, OpenAI, Google, Microsoft, and more). Six weeks, 2–3 hrs/week.
| Track | Duration | Hours/Week | For |
|---|---|---|---|
| Explorer | 8 weeks | 2–3 | Non-technical learners who want to understand AI |
| Practitioner | 10 weeks | 4–6 | Power users who want to use AI tools effectively |
| Builder | 12 weeks | 6–10 | Developers who want to build AI applications |
| Researcher | 16 weeks | 10+ | Advanced learners pursuing deep understanding |
| Your AI Data Footprint (privacy sidecar) | 6 weeks | 2–3 | Recommended alongside any primary track — what gets stored, what trains the model, what stays local |
- Adaptive quiz — 6 questions determine your recommended track
- Validated routing — quiz routing validated against 21 simulated personas (
simulate-personas.ts) - Real resources — every link points to a real paper, course, book, or tool
- Progress tracking — check off completed activities (saved in your browser)
- No account needed — works entirely client-side; nothing leaves your machine unless you explicitly submit the optional outcomes form
- Single HTML file — zero dependencies, works offline
Practicing what the privacy track teaches, the site discloses its own data flow: it has exactly one optional external endpoint — an outcomes form (a 6-week check-in and/or a learning-community interest check) you may choose to submit — and nothing is sent anywhere unless you fill it in and submit it. The community option is an interest check, not a signup for something that exists.
- Stanford HAI AI Index
- 3Blue1Brown neural network series
- Andrej Karpathy's Zero to Hero
- fast.ai Practical Deep Learning
- Anthropic prompt engineering documentation
- HuggingFace NLP Course
- Goodfellow's Deep Learning textbook
- Original papers (Attention Is All You Need, Constitutional AI, Scaling Laws)
- Books (Prediction Machines, Designing ML Systems, Superagency)
Open index.html in a browser. That's it.
Or visit the hosted version: srieg.github.io/ai-curriculum
This curriculum is a single HTML file with zero external dependencies. Fork it, swap in your content, and have a live site in about 20 minutes.
- Click Fork on this repo (top-right on GitHub).
- In your fork, go to Settings → Pages. Set Source to Deploy from a branch, branch
main, folder/ (root). Click Save. - Edit
index.htmlin the GitHub web editor or clone locally. - Your site is live at
https://{your-handle}.github.io/ai-curriculum/within 60 seconds of your first push.
All the content you'll want to change lives in the JavaScript data block (~lines 2200–2720):
| What to edit | Where (approx. line) | Notes |
|---|---|---|
| Assessment questions | const QUESTIONS ~2260 |
6 objects with options[] and per-track scoring weights |
| Track definitions (name, color, duration) | const TRACKS ~2335 |
Keys: explorer, practitioner, builder, researcher, privacy |
| Per-day resources | TRACKS[*].weeks[*].days[*] |
{type, time, title, desc, url, tags[]} |
| Track metadata (philosophy, reading order) | const DEEP_DIVES ~4160 |
Used for context panels, not the day schedule |
| Track display order | const TRACK_ORDER ~2730 |
Simple array of track keys |
| localStorage namespace | 17 keys prefixed ai-curriculum-* |
Must rename in forks — see below |
Before publishing your fork: change the localStorage namespace from ai-curriculum- to your own prefix (e.g., my-ai-path-). If you don't, and a learner visits both the original site and your fork in the same browser, their progress data will collide.
Find-and-replace all 17 keys:
ai-curriculum-theme
ai-curriculum-track
ai-curriculum-progress
ai-curriculum-profile
ai-curriculum-knowledge
ai-curriculum-knowledge-state
ai-curriculum-interests
ai-curriculum-reviews
ai-curriculum-difficulties
ai-curriculum-engagement
ai-curriculum-chat-quality
ai-curriculum-api-key
ai-curriculum-problems
ai-curriculum-mastery-skips
ai-curriculum-cert-name
ai-curriculum-reassessment-dismissed
ai-curriculum-outcomes-optin-dismissed
A single find-and-replace of ai-curriculum- → your-prefix- across the file handles all 17 in one shot.
Attribution is not required under the MIT license, but a link back to the original is appreciated. The community benefit compounds when forks are discoverable.
MIT