What this article covers
- A trustworthy overview of free AI courses, from fundamentals to AI Agents.
- German offerings from KI-Campus and openHPI that often get left out of English lists.
- The logical order in which these courses make sense.
- How to build an actual learning plan from these links instead of just collecting them.
Free AI courses at a glance
The list of truly reputable free courses is short, and that’s a good thing. I’ve focused on official AI websites like Hugging Face and Google/Kaggle, filtering out random link recommendations. These are the providers that deliver substantive content rather than just aggregating links:
| Topic | Link | What you get |
|---|---|---|
| AI Fundamentals | elementsofai.com | The famous introductory course from the University of Helsinki, no math required, available in German |
| Python for AI | kaggle.com/learn/python | The programming language that underpins nearly everything in AI |
| Data Science | kaggle.com/learn | Mini-courses with real datasets, run directly in your browser |
| Machine Learning | kaggle.com/learn/intro-to-machine-learning | The standard ML entry point, using Scikit-learn |
| Deep Learning | kaggle.com/learn/intro-to-deep-learning | Neural networks with Keras |
| Generative AI | kaggle.com/learn | Understanding image and text generation |
| LLMs | huggingface.co/learn | The official course for language models |
| AI Agents | huggingface.co/learn/agents-course | The agents course, currently the most important one |
| NLP | huggingface.co/learn | Natural Language Processing in depth |
| Computer Vision | kaggle.com/learn/computer-vision | Image recognition and CNNs |
| Prompt Engineering | learnprompting.org | Systematic prompt training |
Additions that belong on every list
Three sources stand out because they differ from the rest:
- fast.ai, the most pragmatic deep learning course available, built around practice rather than theory. Free but challenging, suited for anyone who already knows Python.
- DeepLearning.AI Short Courses, Andrew Ng’s team offers concise courses on current topics like RAG, agents, and LLMOps. Each one to two hours long, always up-to-date, completely free.
- Google AI Essentials, Google’s official foundations course for everyday work. Available in German and designed for non-technical learners.
- OpenAI Cookbook and Anthropic Documentation, not courses but primary sources. If you want to understand what models actually do, read the docs instead of summaries.
One key point: Kaggle courses run entirely in your browser with notebooks, while Hugging Face assumes basic Python knowledge. The order listed here works as written: first understand what AI is, then learn the language, then work with the models.
German resources missing from international lists
International compilations often overlook these excellent German offerings:
- KI-Campus (ki-campus.org), the learning platform backed by the Stifterverband foundation, completely free and in German, with courses spanning from fundamentals to specific career fields. The most credible German starting point.
- openHPI (open.hpi.de), the course platform from the Hasso Plattner Institute, free with AI and data courses at university level in German.
- Elements of AI in German, the Helsinki course is also available in German and remains the gentlest introduction.
- German YouTube, channels like Dorflehrer or content from German universities explain AI topics in German, though less systematically.
A sensible progression
| Level | Goal | Course |
|---|---|---|
| 1 | Understand what AI is | Elements of AI or KI-Campus fundamentals |
| 2 | Learn to code | Kaggle Python |
| 3 | Machine learning | Kaggle Intro to ML |
| 4 | Go deeper | Hugging Face LLM or Agents course |
| 5 | Specialize | Computer Vision, NLP, or Prompt Engineering |
If you don’t want to code, stop at Level 1 plus Prompt Engineering and you’ll still have the knowledge that matters in the workplace. If coding is your goal, Python comes before everything else, since all Kaggle and Hugging Face courses build on it.
The difference between collecting and learning
Nearly every link list fails at the same point: you save it and never open it again. What actually works is taking one course at a time, at your own pace, with a fixed learning plan. Create a learning plan with AI shows how to build one yourself, and if you want to retain what you learn, flashcards with AI is the logical next step. For German learners, start with KI-Campus or Elements of AI in German, then move into the Kaggle sequence.
Frequently asked questions about free AI learning
Are these courses really free?
Yes, every one listed is completely free. Kaggle and Hugging Face are free by default, Elements of AI is free as a university project, and KI-Campus and openHPI are publicly funded educational platforms.What’s the best course to start with?
Elements of AI if you want understanding without programming, or KI-Campus in German for the German context. If you want to start coding immediately, begin with the Kaggle Python course.Do I need math for these AI courses?
Not for Elements of AI or KI-Campus. For the deeper Kaggle and Hugging Face courses, yes, you’ll need some foundational understanding. Our article on learning math with AI can help.Are the courses available in German?
Elements of AI yes, KI-Campus and openHPI absolutely. Kaggle and Hugging Face run in English, but the language is technical and manageable, and translation tools help if you need them.How long does a course take?
Elements of AI is structured for six weeks part-time, Kaggle mini-courses each take four to six hours, and Hugging Face courses span several weeks. The full list represents about half a year of work, not a weekend.Do I get a certificate?
Elements of AI, KI-Campus, and openHPI offer certificates, some free and some for a small fee. Kaggle provides completion certificates. In the job market, what matters more is what you can do than the paper, and you show that through your own projects.Are free courses enough to get a job?
Absolutely for learning and understanding, but changing careers requires building your own projects you can show. The courses are the foundation, not the replacement.What’s the difference between Kaggle and Hugging Face?
Kaggle is the learning platform with in-browser notebooks, Hugging Face is the platform for models themselves. Kaggle teaches the basics, Hugging Face connects you to real models.Is Prompt Engineering worth taking as a course?
Yes, and learnprompting.org is the best free resource for it. It’s the only course in this list that requires no programming, which makes it the most practical entry point for many people.Are there free German courses for companies?
KI-Campus has specific learning paths for professions and companies, and openHPI offers courses for decision-makers. Both are the credible German foundation before moving to English content.What’s the right order to learn AI?
First understand (Elements of AI or KI-Campus), then code (Kaggle Python), then model (ML, deep learning), then specialize (LLMs, agents, computer vision). If you don’t want to code, stick with understanding and prompt engineering.Is Hugging Face too hard for beginners?
The Learn courses do require Python beforehand, yes. If you don’t know Python yet, take the Kaggle course first, otherwise the Hugging Face entry point will frustrate quickly.What’s the advantage of German offerings?
KI-Campus and openHPI teach in German, removing the language barrier, and they often have more direct relevance to the German job market and local regulations.Can I take these courses on my phone?
Elements of AI and KI-Campus yes, they’re designed for it. Kaggle and Hugging Face require a computer because you write code yourself, which is tedious without a keyboard.How do I stay motivated when learning alone?
A fixed weekly schedule, a small project as a goal, and a system to track progress. Create a learning plan with AI explains the setup, and AI itself can help keep you on track.Free doesn’t mean worthless
Every course here is better than most things people pay for. Elements of AI for understanding, Kaggle for doing, Hugging Face for depth, KI-Campus and openHPI for the German framework. The only real cost is time, and anyone who completes one of these instead of collecting them all has learned more than most people who buy three courses and finish none.