Eitan Turok

A friend recently asked me: if you could start any startup right now, what would it be?

Intro I thought it would be fun to go through How to Scale Your Model, a small book on how to train and perform inference on LLMs at scale.

Intro With Claude’s help, I wrote ~30 interview questions on LLM Inference, focusing on the systems side of things like FLOPs, attention, and the kv-cache.

First let’s define our parameters: matrices $A \in \mathbb{R}^{m \times n}$; $B \in \mathbb{R}^{m \times n}$; $C \in \mathbb{R}^{p \times n}$; $W \in \mathbb{R}^{m \times m} = \text{diag}{w_1, …, w_m }$ where $w_i > 0$ vectors $b \in \mathbb{R}^n$; $d \in…

Introduction Dynamic programming (DP) is one of the most common algorithmic paradigms, used everywhere from DNA sequencing and financial modeling to matrix multiplication and shortest path algorithms.