Learn Inference: an interactive guide to serving generative models in production

Training teaches a model what it knows. Inference is everything that happens afterward, every time somebody uses it, and it is where the bill actually lands. Serving a generative model well means working across a strange range of the stack: attention kernels at one end, GPU procurement across three clouds at the other.

This is an interactive companion to Inference Engineering by Philip Kiely. It follows the book’s structure and covers the same ground. I rewrote the explanations and built simulators for the parts you grasp faster by turning a dial than by reading a paragraph.