Free, hands-on AI courses that run anywhere, from the team building Synapsa, an AI-native learning platform.

Every lesson here is a notebook you do, not one you read. You fill in the exercises, and each one checks itself the moment you run it. Hints stay hidden until you ask for them. Worked solutions and the full autograded rubric, with partial credit and an official score, come with free enrolment on Synapsa, opening in the coming weeks. Some lessons are C or C++ exercises, built and graded the same way. Open any of them in Google Colab, Kaggle, Binder, GitHub Codespaces or your own Jupyter: the first cell installs what the lesson needs at the versions it was measured with, and does nothing where it is already there.

The courses are aimed at the work people are actually paid for: evidence an inspector can re-run, a model a risk committee can sign off, an alarm a plant manager will trust. Every number a lesson prints is computed by code you run, and every claim it makes about the world cites a primary source.

Where to start

If you want to…	Course	First lesson
Show that an AI system meets the EU AI Act, with evidence an inspector can re-run	programmes/ai-act-conformity/	Colab
Validate a model well enough to sign it off: calibration, challengers, explainability, the committee pack	programmes/model-risk/	Colab
Take predictive maintenance from sensor physics to an alarm threshold priced in money	programmes/predictive-maintenance/	Colab
Pull fields out of invoices and contracts, and prove how often you are wrong	programmes/document-intelligence/	Colab
Make a simulated humanoid stand and walk, then measure how far that is from a robot	flagships/humanoid-lab/	Colab
Build search and RAG you can measure: retrieval, chunking and the context budget	lessons/	Colab
Start from nothing: the 8 GB machine, tokenisers from scratch, a language model on a CPU	lessons/	Colab

The model-risk, predictive-maintenance and document-intelligence courses build on lessons/T00-L01-the-8gb-track, which makes the profiler and the compute tiers every lesson is held to; start there if you have not done it. Each programme's MODULES.md is its map and shows which lessons exist and which are still only specified; the humanoid lab's is its README.md, and lessons/ has its own.

What exists today

54 lessons, every one passing all 14 gates in QUALITY.md and independently reviewed. 7 are compiled C or C++ exercises. Every notebook opens in Colab, Kaggle, Binder, Codespaces or local Jupyter. Every MODULES.md marks a module built only when its lesson exists and has passed; everything else says specified, and means it.

Area	Built	Specified, not built
flagships/humanoid-lab	8 lessons + capstone	—
lessons/ — track lessons	7	—
programmes/ai-act-conformity	9 of 9	0
programmes/document-intelligence	11 of 11	0
programmes/model-risk	10 of 10	0
programmes/predictive-maintenance	9 of 9	0

What's coming

A new drop of lessons every week. Watch or star this repository to hear about each one as it lands.

timeline
    September 2026 : Document intelligence for regulated operations, and retrieval and RAG begins
    October 2026 : AI agents and the protocols they speak, AI for science, serving models on modest hardware, post-training from scratch, and evaluating across languages
    November 2026 : A new industry programme on evidence-based credit decisions
    December 2026 : Deeper courses in the industries we already cover

How a lesson works

Press Run all before you have written a line. Every cell still runs: each unfinished exercise says it is not implemented yet, and the notebook ends with a progress board, one line per exercise, marked ✅ passed, ❌ failed or ⏳ not started. Fill in an exercise and re-run its cell, and its check tells you whether you are right, and if not, what the likely mistake is. Each exercise has two hints: the first says what to think about, the second the approach in words. Neither gives you the answer. Your score against the full rubric, and the worked solution, come with enrolment on Synapsa.

Anything a lesson measures on your machine, like a time or a memory peak, prints with its unit and will differ from ours. Every other number a finished lesson prints is identical on Python 3.11 and 3.12; every lesson is checked for that before it is released. The workflow above checks on Linux, which is what Colab, Kaggle, Binder and Codespaces run, that every notebook opens and runs.

On Kaggle, a lesson that installs a package or fetches a file needs Internet switched on in the notebook's settings, and Kaggle allows that only for phone-verified accounts. Colab, Binder and Codespaces need nothing.

To work locally instead, with Python 3.11 or 3.12:

python -m venv .venv && .venv/bin/pip install -r requirements.txt
.venv/bin/jupyter lab

On Windows the two commands are .venv\Scripts\pip and .venv\Scripts\jupyter.

The bar every lesson clears

QUALITY.md sets out 14 gates. The short version:

• measurable objectives, and prerequisites that are real lessons;
• scaffolded exercises with instant feedback, an autograded rubric, a worked solution and a self-check;
• runs top to bottom in a fresh interpreter, inside a declared compute tier and time budget;
• data that is free and needs no registration, named with its licence;
• every claim about the world cites a primary source, and no measured or computed number in the prose is typed by hand;
• the notebook opens and runs alone on Python 3.11 and 3.12, and Run all is a good experience.

A lesson is marked built only after an independent reviewer has tried to break it, including by grading plausible wrong answers against its rubric.

For maintainers

• lessons/, flagships/, programmes/: track lessons, flagship subtrees, industry programmes.
• Lessons are written and gated in a separate source that holds the worked solutions and the rubrics; this repository receives the student version of each lesson once it passes.
• tools/notebooks.py generates each lesson.ipynb and its launcher cell (--check for drift), and tools/verify_portable.py runs each notebook alone in a minimal kernel on Python 3.11 and 3.12. .github/workflows/lessons.yml runs both on Linux on every change, and fails if a solution or rubric file is ever committed here.
• brand/ holds the Synapsa Commons lockups used here and at the top of every notebook.

The 2017-2025 tutorials

From 2017 to 2025 this repository carried a curated collection of machine-learning and deep-learning tutorials, much of it gathered from other people's work. It helped a great many learners, and we are grateful to everyone who starred it, forked it and contributed to it. That collection has now been retired from public view, so that everything published here is original work we can stand behind and license. Synapsa Commons replaces it.

Licence

Synapsa Commons is © 2026 RealAI, free to learn from and share, and licensed under CC BY-NC-SA 4.0; the legal code is in LICENSE and a plain-words summary in NOTICE.

You may	You must	You may not
use the lessons to learn and to teach; copy, fork, share and adapt them	credit "Synapsa Commons by RealAI" with a link here, say what you changed, and share adaptations under the same licence	use them, or anything adapted from them, for commercial advantage or payment: selling them, charging for a course or training built on them, or packaging them into a paid product or service

Commercial licences are available from RealAI. Third-party material inside a lesson keeps its own licence, named in that lesson's assets/SOURCE.md or claims.yaml. The Synapsa name and logo belong to RealAI and are not licensed.