Downloads and runnable resources

Choose the complete source when you want to run or adapt the course. A notebook on its own does not install the shared package, dependencies, or data.

Complete project

Download the complete source ZIP. Extract it, keep its folder structure, and follow first-run setup. It includes the Python package, all module scripts/notebooks, teaching pages, contracts, fixtures, tests, and authoring templates. If a download link does not open, use the offline access route.

The GitHub links on each reading page identify its exact source revision. Use a tested commit or semester edition when adapting.

Notebooks

Start with a clean source notebook when working through the steps. Use the executed copy to inspect saved outputs without rerunning. Executed copies come from the actual verified CI kernels, not invented results. Their saved presentation still requires manual review.

Clean source and executed notebook copies by module
Module	Start fresh	Inspect saved outputs
1. ML and AI foundations for practical decisions	Source notebook, module 1	Executed notebook, module 1
2. Python for reproducible ML workflows	Source notebook, module 2	Executed notebook, module 2
3. Data processing and trustworthy inputs	Source notebook, module 3	Executed notebook, module 3
4. Exploration and visual communication	Source notebook, module 4	Executed notebook, module 4
5. Regression and reliable evaluation	Source notebook, module 5	Executed notebook, module 5
6. Classification and decision trees	Source notebook, module 6	Executed notebook, module 6
7. Clustering and learning paradigms	Source notebook, module 7	Executed notebook, module 7
8. Neural networks with TensorFlow	Source notebook, module 8	Executed notebook, module 8
9. Neural NLP for real messages	Source notebook, module 9	Executed notebook, module 9
10. LLMs and evaluated generative workflows	Source notebook, module 10	Executed notebook, module 10
11. Fine-tuning language models responsibly	Source notebook, module 11	Executed notebook, module 11
12. Deployment and monitoring	Source notebook, module 12	Executed notebook, module 12
13. Ethics and fairness across the lifecycle	Source notebook, module 13	Executed notebook, module 13
14. Bounded agents and AI workflows	Source notebook, module 14	Executed notebook, module 14

On GitHub, the source notebooks are in notebooks/lessons/. The built site supplies the individual download links above. Keep a notebook inside the complete source tree when running locally.

Practice with less support

Modules 5, 6 and 8 provide separate starter notebooks and worked solutions. Start with a learner function, use its visible feedback, then inspect solutions after attempting the task. Sources are also under notebooks/practice/ in the complete archive. Executed starters retain unfinished feedback; executed solutions contain the actual development sandbox outputs. Those are distinct evidence claims.

Reviewed external data caches

Initial data downloads require internet. These separately attributed caches preserve the reviewed original archive/member bytes and provenance. Follow the module's cache instructions rather than scattering extracted files into its source.

Read source rights and provenance before reuse. Public messages and historic demographic labels have privacy and interpretation limits. Model weights are not bundled in the source ZIP.

Choose an execution environment

Modules 1–2 can start with Python alone. The core setup supports data processing and classical ML. TensorFlow, local language models, and the app use the separate framework environment, currently verified on Linux CPU.

Use scripts and text/table reports when they fit your access needs. Reading a saved result is useful preparation but does not itself demonstrate implementation.

View source for this page on GitHub