AI Engineer
How models really learn
AI Engineer is a free course on how AI models really work inside. It explains neural networks, how training adjusts a model with gradient descent, what transformers and attention do, how fine-tuning changes a model, and how engineers measure whether a model is any good.
What you'll learn
- Explain how a neural network turns inputs into predictions
- Describe training, loss and gradient descent clearly
- Understand transformers and attention at an intuitive level
- Know when fine-tuning helps and when prompting or RAG is better
- Read model results critically: overfitting, bias and evals
Who this course is for
- Computer science students going deeper into AI
- Developers moving into machine learning
- Anyone preparing for AI and ML interviews
Inside the first unit
Neurons and weights, loss and gradient descent, tokens and embeddings.
- Neurons and weights
- Loss and gradient descent
- Tokens and embeddings
- A neuron computes a weighted sum and passes it through an activation.
- Training repeats: predict, measure the loss, follow the gradient downhill.
- Text becomes tokens; meaning becomes vectors you can compare.
After the first unit, Pli keeps going: new units are written for your level and goals as you learn, and every question is checked by a second AI before you see it.
Key terms in this course
Questions about AI Engineer
How much maths is involved?
Lessons build intuition first and introduce the maths only where it helps. You don't need to be a maths expert to follow along.
Will this help with interviews?
Yes. The concepts covered are the ones that come up in AI and ML interviews, explained so you can say them in your own words.
Is it free?
Yes, AI Engineer is one of the free core courses.
Free, five minutes a day. No sign-up needed to try your first lesson.
Updated 18 October 2026