--- format: html: title: "Introduction to Machine Learning Systems" date: today date-format: long doi: "v0.7.2" doi-title: "Version" author: name: Vijay Janapa Reddi email: vj@eecs.harvard.edu url: https://vijay.seas.harvard.edu affiliation: Harvard University --- ::: {.content-visible when-format="html:js"} ```{=html}

Welcome

Machine learning has evolved from a research discipline into an engineering practice. Building systems that learn from data requires more than understanding algorithms; it demands expertise spanning data pipelines, model development, deployment constraints, accelerator behavior, and operational practice. AI capability is not a software feature; it is an emergent property of the co-design of Data, Algorithm, and Machine under both statistical and computational constraints. This book introduces AI engineering as the discipline of building intelligent systems that operate in the real world, then develops its ML systems foundation at the node level. The treatment covers four areas: foundations (system characteristics, development workflows), building (deep learning mathematics, architectures, framework internals), optimization (compression, hardware acceleration, benchmarking), and deployment (serving infrastructure, operations, responsible engineering). The emphasis throughout is on the quantitative trade-offs among data, algorithms, and machines.

Machine Learning Systems Book Cover

Introduction to Machine Learning Systems

Publisher: The MIT Press (2026)

📖 Download PDF
```

Learning Path

The book progresses through four stages: - **Part I: Foundations**: Establish the conceptual foundation and mental models that underpin effective systems work. - **Part II: Build**: Engineer complete workflows from data pipelines through training infrastructure. - **Part III: Optimize**: Transform theoretical understanding into systems that run efficiently in resource-constrained environments. - **Part IV: Deploy**: Navigate serving, operations, and responsible engineering practices.

Prerequisites

This book assumes: - **Programming proficiency** in Python with familiarity in NumPy - **Mathematics foundations** in linear algebra, calculus, and probability at the undergraduate level - Prior ML experience is helpful but not required; @sec-neural-computation provides essential background

Support Our Mission

```{=html}

2026 Goal: Help 100,000 students learn ML Systems. Sponsors like the EDGE AI Foundation match every star with funding that supports learning.

Loading... ⭐ Star on GitHub

Support us on Open Collective →

``` ```{=html} ```

Listen to the Machine Learning Systems Podcast

```{=html}

This short podcast, created with Google's NotebookLM and inspired by insights from our IEEE education viewpoint paper, offers an accessible overview of the book's key ideas and themes.

```

Contributing

::: {.content-visible when-format="html"} This is a collaborative project. The project [contribution guidelines](https://github.com/harvard-edge/cs249r_book/blob/main/book/docs/CONTRIBUTING.md) explain how to participate. Feedback, corrections, and new ideas are welcome through GitHub [issues](https://github.com/harvard-edge/cs249r_book/issues). ::: ::: {.content-visible when-format="pdf"} This is a collaborative project. The project contribution guidelines explain how to participate, and feedback, corrections, and new ideas are welcome through the GitHub issue tracker at the book's repository (`harvard-edge/cs249r_book`). ::: :::