What turns a language model into an agent: a reasoning loop, tools it can call, and a goal it works toward across several steps instead of one reply — plus a tour of the frameworks (LangGraph, AutoGen, and others) built to orchestrate them.
Start reading →A hands-on companion to Agentic AI — build a tool-calling agent from scratch around a small open model, runnable and editable in a Colab notebook.
Open in Colab →A 16-part intuition-first tour of how GPT-style models work — from prediction to next-token generation, then to the hardware it runs on. Pair it with the companion AI Core Math Review series if the underlying math is rusty, and Build Your Own Transformer for a hands-on notebook.
Start reading →A hands-on companion to Understanding Transformers — assemble a working transformer yourself in a guided Colab notebook, in the order you'd actually build one rather than the order that's easiest to learn.
Open in Colab →The core math principles neural networks lean on most: scalars, vectors, dot products, and matrices — the toolbox behind attention, embeddings, convolutions, and everything in between.
Start reading →A hands-on companion to AI Core Math Review — the same scalar, vector, dot product, and matrix examples, runnable and editable in a Colab notebook.
Open in Colab →