Elements of AI: Building AI
University of Helsinki & MinnaLearn
Last updated June 29, 2026
Elements of AI: Building AI is the second course in the Elements of AI series from the University of Helsinki and MinnaLearn — the follow-on to Introduction to AI. Where the first course explains what AI is, this one opens the hood: it introduces the methods that actually make AI work — optimization, probability and Bayes' rule, machine learning (linear regression, nearest neighbor, overfitting), and neural networks through to deep learning. It's built around three difficulty levels you can move between freely, from multiple-choice exercises to writing Python, so you can go as deep as your programming comfort allows. You finish by shaping your own AI idea and sharing it with the course community. The course is free to take; an electronic certificate is available to purchase on completion. Plan for roughly 50 hours at your own pace.
What you'll learn
- Optimization and search (hill climbing) as the basis of AI problem-solving
- Reasoning under uncertainty with probability and Bayes' rule
- Core machine learning: linear regression, nearest neighbor, and avoiding overfitting
- Neural networks, from logistic regression through to deep learning
- Framing and presenting your own AI idea
Frequently asked questions about Elements of AI
Who is Elements of AI for?
Anyone who has grasped the basics of AI and wants to understand the methods behind it — machine learning and neural networks — at whatever depth their programming background allows.
Is Elements of AI free?
Yes — Elements of AI is completely free to take.
What are the prerequisites for Elements of AI?
None formally; Introduction to AI (Part I) is the natural lead-in. Some basic Python is recommended for the hands-on track, but the course can be completed without programming via its lower difficulty levels.
Does Elements of AI offer a certificate?
Yes. Optional electronic certificate available to purchase on successful completion.
Why we suggest this course
The natural next step after a conceptual intro: it bridges "I understand what AI is" and "I see how AI is actually built," without demanding a maths or programming background up front. The three-level design is the draw — a non-programmer can follow the ideas through multiple-choice exercises, while anyone comfortable with Python can go hands-on with the same material. One thing to know: the course is free, but the certificate is a paid optional extra.