AI Foundations

9 essential concepts that power modern AI — from tokenization to LoRA. Interactive diagrams, real-world examples, and knowledge checks.

0/9

The process of converting raw text into numerical tokens that neural networks can understand.

How LLMs generate text one token at a time by sampling from probability distributions.

Crafting effective inputs to guide LLM behavior without retraining the model.

Systems that extend LLMs with actions, tools, and memory to complete complex tasks autonomously.

Retrieval Augmented Generation — enhancing LLMs with external knowledge to reduce hallucinations.

Reinforcement Learning from Human Feedback — aligning AI with human preferences for helpful, safe output.

A generative model that learns compressed representations for creating new data.

Generating images and video by learning to reverse a noise-adding process.

Low-Rank Adaptation — efficient fine-tuning without retraining billions of parameters.