Machine Learning Systems

/`<br>and then relativizes any navbar.href value that lives inside the<br>site-url's domain. The shared `href: https://mlsysbook.ai/` in<br>navbar-common.yml therefore gets emitted as `./index.html`, which on<br>vol1/vol2/tinytorch/etc. resolves to *that subsite's* root rather than<br>the ecosystem landing page. Override the brand link href in JS after<br>Quarto's render: this is a one-line user-experience fix that does not<br>fight Quarto's URL machinery.<br>-->

📖 ML Systems — an open-access textbook on the engineering of intelligent systems. Vol I: Foundations → · Vol II: At Scale →<br>🛠️ Alongside the book: TinyTorch (build) · Hardware Kits (deploy) · MLSys·im (model) · Labs (explore) · StaffML (practice)<br>📬 Newsletter: ML Systems insights & updates — Subscribe →

TWO-VOLUME TEXTBOOK

Machine Learning<br>Systems.

The physics of AI engineering.

A rigorous, principles-first treatment of how ML systems are built, optimized, and deployed — from a single machine to fleet-scale infrastructure.

Harvard University · MIT Press 2026

Actively maintained<br>Last updated April 2026<br>Release notes

Volume I

Introduction to Machine Learning Systems

Volume I downloads:<br>HTML<br>PDF<br>EPUB

Volume II

Machine Learning Systems at Scale

Volume II downloads:<br>HTML<br>PDF<br>EPUB

Explore the Curriculum

A complete curriculum for AI engineering.

Choose a path: read the books, explore trade-offs in labs, build the internals with TinyTorch, model constraints with MLSys·im, deploy on real hardware, practice with StaffML, or adopt the full course with the Blueprint.

For Students & Learners

EXPLORE

Labs

Interactive Marimo notebooks. Change a parameter, see what breaks, build intuition.

Lab 15 · Sustainable AI

Explore

BUILD

TinyTorch

Build your own ML framework from scratch across 20 progressive modules. Zero magic.

tinytorch — tensor.py

class Tensor:<br>def __init__(self, data):

self.data = data<br>self.grad = 0.0<br>self._backward = lambda: None

MODEL

MLSys·im

First-principles performance modeling. One command, every bottleneck.

$ mlsysim eval Llama3_70B H100 --batch-size 1

mem-bound<br>compute-bound

b=1<br>b=32<br>b=128

Arithmetic Intensity<br>FLOP/s

DEPLOY

Hardware Kits

Deploy ML to Arduino, Seeed, Grove, and Raspberry Pi. Real memory limits, real power budgets.

Arduino · Seeed · Grove · Raspberry Pi

For Career & Instructors

PRACTICE

StaffML

Physics-grounded interview questions for ML systems roles. Vault, drills, and mock interviews.

Systems Design<br>L5 · Staff

A 70B model needs 1,000 req/s.<br>Walk through your hardware selection<br>and parallelism strategy.

Hardware

Parallelism

Trade-offs

Cloud

Edge

Mobile

TinyML

ADOPT

Instructor Hub

The AI Engineering Blueprint: two-semester syllabi, pedagogy guide, rubrics, and TA handbook.

The Blueprint — Course Architecture<br>ML Systems · Two-Semester Curriculum

Semester 1: Foundations<br>16 wks · Vol I · 8 assignments

Semester 2: At Scale<br>16 wks · Vol II · capstone

Assessment<br>Rubrics · Peer review · Grading

Teaching Staff<br>Pedagogy · TA handbook

READY

TEACH

Lecture Slides

35 Beamer decks with speaker notes and 266 original SVG diagrams. Drop in and teach.

Intro<br>Systems<br>DNN

Training<br>Accel<br>Deploy<br>Ethics

The Iron Law of ML Systems

T = D/BW + O/(R·η) + L

Data Term — memory bandwidth

Compute Term — utilization η ≤ 0.7

Latency Term — orchestration overhead

Harvard University · ML Systems<br>12 / 38

FOLLOW

Newsletter

Updates on the curriculum, new chapters, and what the community is building.

MLSysBook Weekly

New: Vol II Ch. 14 — Fault Tolerance

Updated: TinyTorch Module 12

Community: 500+ PRs merged

Milestone: 23,000 GitHub stars

Join 12,000+ subscribers

Support the Mission

OUR MISSION

AI education should be<br>free and open to everyone.

Everyone calls AI the new electricity — but electricity is useless without engineers who can build the grid. For AI to be efficient, reliable, and safe, the world needs engineers who understand how to build it.

That knowledge should be accessible to anyone willing to learn. This curriculum is our commitment to making it so.

Live readership — 180+ countries

23,000+ stars · 243,000+ readers · 180+ countries

Our goal: 1,000,000 AI engineers by 2030

Next milestone: 100,000 ★ — we're at 23,000+.<br>Every star helps others discover this resource.

Star on GitHub

Meet our sponsors & partners →