Choose a useful project.

Six existing foundations plus thirty new project courses. For people across regions underserved or excluded by mainstream providers—not one country. All original lessons, fixtures and handbooks are free to download. Hosted services have their own eligibility, privacy and quota conditions.

Advanced prerequisites are labelled. No local AI model download, payment platform or account signup is required to read. A manual reference exercise is not a live model run.

Foundations

• AI, without the guesswork — Foundation

Everyday work

• Your first business assistant — Foundation

Data & spreadsheets

• Make messy data useful — Foundation

Build with AI

• From idea to working page — Foundation

Automation

• Automate the boring bit—safely — Foundation

Open AI systems

• Understand the AI plumbing — Foundation

Cloud & code

• A cloud lab someone else can rerun — Coding onramp
• Turn records into a tested Python report — Coding onramp
• Measure an algorithm instead of guessing — Python prerequisite

Creative production

• Direct an illustration with a measurable brief — Open entry
• Compare generations without moving the goalposts — Open entry

Measured machine learning

• Build a prediction and an honest baseline — Python + statistics
• Augment training data without contaminating the test — Python + ML prerequisite
• Classify feedback with an uncertainty bucket — Python onramp
• Fit a regression and inspect its residuals — Python + statistics
• Build a preprocessing step that cannot see the future — Python + ML prerequisite
• Use a pipeline and a genuinely held-out set — Python + statistics
• Find clusters—and show why they are not facts — Python + statistics

Assistants & APIs

• Make a document assistant that knows when to stop — Cloud API onramp
• Build an FAQ assistant with a human exit — Cloud API onramp
• Connect an AI API without trusting every response — JavaScript / Python prerequisite

AI foundations

• Test one AI claim in the real world — Open entry
• Audit a tool with evidence, not an ethics slogan — Open entry

Data & evidence

• Build a report that reconciles to its source — Spreadsheet onramp
• Repair dirty records without inventing facts — Spreadsheet onramp
• Join two datasets without multiplying the truth — Python prerequisite
• Make cleaning decisions testable with Pandas — Python + Pandas prerequisite
• Make charts that tell the checked truth — Python onramp

Planning & uncertainty

• Plan a timetable that obeys its constraints — Logic onramp
• See a gradient step succeed—and fail — Algebra prerequisite
• Report uncertainty instead of fake precision — Basic statistics
• Make a reversible stock-planning decision — Spreadsheet onramp
• Build a small project plan you can verify — Open entry
• Measure whether a workflow actually saves time — Open entry

Make Mula

• Create a campaign pack without invented performance — Open entry
• Make Mula: useful work, honest offers — Portfolio capstone

Published scope

Thirty strong-fit topics were built as original courses. Eleven supporting inputs are embedded inside related lessons. The twenty-seven conditional and six held topics were not published. No affiliation, accreditation or copied course material is implied.