Home
/ AI System Design
/ Foundations How AI Systems Differ Topic 3 lessons ~2h
Lessons, in order 01
What a Model Changes Put a language model in the request path and five assumptions of ordinary system design break. The component is slow (seconds, not milliseconds), metered by size (tokens, not requests), … Foundations
30 min
02
The Design Method Every design in this course is worked in the same six steps: requirements → numbers → architecture → failure → cost → security. The order matters. Numbers before boxes, because token … Foundations
45 min
03
The 2026 Landscape An AI system in 2026 is seven layers, and each layer has settled on a small number of real options. From the bottom: compute, cluster, serving, gateway, data and context, orchestration, and … Foundations
45 min
About this topic A model in the request path changes what “slow”, “expensive”, “correct” and “trusted” mean.
This topic establishes those changes, gives you one method for designing around them, and maps
the 2026 tool landscape so that every later lesson has somewhere to hang.
# Lesson The question it answers 01 What a Model Changes Which assumptions of ordinary system design stop holding? 02 The Design Method In what order do I work a design, and which numbers come first? 03 The 2026 Landscape Which layers exist, what fills each, and which standards connect them?
If you have designed ordinary backends before, lesson 01 is the one that recalibrates you.