Learn
Hands-on AI and data courses built around the datasets and models on the hub — plus the courses from other providers that people most often recommend.
139 courses
A method for choosing between tools that survives being questioned a year later.
Definitions, the disagreements behind them, and why the word carries so much argument.
A realistic account of the capabilities, with the failure mode for each.
Finding the use that pays, and recognising the ones that will fail.
Matching a store to a workload, and why "it depends" has a small number of actual answers.
What has to exist around a model before it can serve real traffic, and the order to build it in.
The pipeline end to end — index, retrieve, assemble, generate — and where each stage fails.
The shape of a working pipeline — sources, landing, transformation, serving — and the decisions at each boundary.
The loop, the tools and the stopping condition — and when a plain pipeline is the better answer.
What actually changes a model's output, in order of effect — and the habits that waste effort.
The problem the protocol solves, its three primitives, and when to use it instead of a plain API call.
What it means for one agent to delegate to another, and the problems that appear immediately.