What do you think?
Rate this book
Building Intelligent Systems: A Guide to Machine Learning Engineering
Produce a fully functioning Intelligent System that leverages machine learning and data from user interactions to improve over time and achieve success.
This book teaches you how to build an Intelligent System from end to end and leverage machine learning in practice. You will understand how to apply your existing skills in software engineering, data science, machine learning, management, and program management to produce working systems. Building Intelligent Systems is based on more than a decade of experience building Internet-scale Intelligent Systems that have hundreds of millions of user interactions per day in some of the largest and most important software systems in the world.
What You'll Learn
Understand the concept of an Intelligent System: What it is good for, when you need one, and how to set it up for success
Design an intelligent user experience: Produce data to help make the Intelligent System better over time
Implement an Intelligent System: Execute, manage, and measure Intelligent Systems in practice
Create intelligence: Use different approaches, including machine learning
Orchestrate an Intelligent System: Bring the parts together throughout its life cycle and achieve the impact you want
Who This Book Is For
Software engineers, machine learning practitioners, and technical managers who want to build effective intelligent systems
This book teaches you how to build an Intelligent System from end to end and leverage machine learning in practice. You will understand how to apply your existing skills in software engineering, data science, machine learning, management, and program management to produce working systems. Building Intelligent Systems is based on more than a decade of experience building Internet-scale Intelligent Systems that have hundreds of millions of user interactions per day in some of the largest and most important software systems in the world.
What You'll Learn
Understand the concept of an Intelligent System: What it is good for, when you need one, and how to set it up for success
Design an intelligent user experience: Produce data to help make the Intelligent System better over time
Implement an Intelligent System: Execute, manage, and measure Intelligent Systems in practice
Create intelligence: Use different approaches, including machine learning
Orchestrate an Intelligent System: Bring the parts together throughout its life cycle and achieve the impact you want
Who This Book Is For
Software engineers, machine learning practitioners, and technical managers who want to build effective intelligent systems
365 pages, Paperback
Published March 7, 2018
17 people are currently reading
97 people want to read
About the author
Geoff Hulten
1 book4 followersGeoff Hulten is a PhD in machine learning and author of Building Intelligent Systems. He has managed applied machine learning projects for over a decade, working on dozens of Internet scale machine learning efforts. He teaches graduate level machine learning at the University of Washington. His research has appeared in top international conferences, received thousands of citations, and won a SIGKDD Test of Time award for influential contributions to the data mining research community that have stood the test of time.
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 6 of 6 reviews
July 12, 2018
As a Machine Learning scientists working at a large software engineering company, I strongly feel like this book should be one of the mandatory readings for anybody working on real-world machine learning systems, regardless of their role (software engineer, data scientist, product manager, etc.). Most Machine Learning books will teach you the various ways you can train a model, what loss function to use, how to avoid overfitting, etc, but none of them will explain to you the most important part: how to ship your model to have customer impact, and how to design your user experience such that every user interaction improves your sytem. This book explains it in great details, with lots of concrete examples and very clear explanations about each and every steps required to deliver a successful intelligent experience. The summary at the end of each chapter are very useful when coming back to the book later to review a section! I also really liked the tone of the book, full of funny examples and witty remarks. Way to make a technical read entertaining!
August 25, 2019
I'm about as lay as a layman can get when it comes to understanding concepts like AI and telemetry. At the opposite end of the spectrum are people like my brother, computer scientists who can stare at screens (displaying what might as well be hieroglyphics to a cromagnon like me), working on projects I'm afraid to ask about, because I know in advance I'm not going to grasp more than a tenth of what is being conveyed.
I read this book, and my brother read it, and we both got a lot out of it. Anything that accessible and yet useful has more than achieved its goals.
Geoff Hulten explains in clear, concise language, the options that one has in creating methods for machines to learn and improve both their knowledge and capacity for learning. When presenting each method, he lists the various drawbacks and advantages of each system and (most importantly) allows that he doesn't have all the answers, and that there is always a place for innovation (provided the innovator is prepared to do a lot of debugging, second-guessing, and deep thinking about what could go wrong, from a machine taking a command too literally or bad actors trying to profit from your hard work by foisting unwanted advertisements on users of your machine, which, if they become invasive enough, may start costing you enough money and time to destroy your business).
There are some charts and graphs, but not many. The book may not be granular enough in explicating certain concepts or citing tactics for machine learning improvement for those who are already very familiar with these concepts and well in advance of even others in their fields where information is gold and telemetry the coin of the realm, but I suspect that even the most capable person in the field could use the book as a refresher, while it more than served my purposes as a primer. Highest recommendation.
I read this book, and my brother read it, and we both got a lot out of it. Anything that accessible and yet useful has more than achieved its goals.
Geoff Hulten explains in clear, concise language, the options that one has in creating methods for machines to learn and improve both their knowledge and capacity for learning. When presenting each method, he lists the various drawbacks and advantages of each system and (most importantly) allows that he doesn't have all the answers, and that there is always a place for innovation (provided the innovator is prepared to do a lot of debugging, second-guessing, and deep thinking about what could go wrong, from a machine taking a command too literally or bad actors trying to profit from your hard work by foisting unwanted advertisements on users of your machine, which, if they become invasive enough, may start costing you enough money and time to destroy your business).
There are some charts and graphs, but not many. The book may not be granular enough in explicating certain concepts or citing tactics for machine learning improvement for those who are already very familiar with these concepts and well in advance of even others in their fields where information is gold and telemetry the coin of the realm, but I suspect that even the most capable person in the field could use the book as a refresher, while it more than served my purposes as a primer. Highest recommendation.
February 9, 2019
Great book that goes through the design philosophy of engineering intelligent systems (system of systems of machine learners stitched together by meta-algorithms, heuristics, or other machine learners). The first use case is a bit trivial (a smart toaster), but sufficient as an illustration on how to gather end-user telemetry and think about how one can anticipate for sudden changes in data distribution that would impact model performance. This is similar to learning regime change in high-frequency trading algorithms it seems. Also goes over a lot of fundamentals that are a useful review, especially in the domain of classification. Enjoyable read. Highly recommended
November 9, 2024
Good intro, thats all, nothing too crazy
July 12, 2021
If you consider as highest point of your Machine Learning career most effective toaster selling - this is book for you. You'll be most competent engineer with best sales pitches by far. Your marketing will take you on every meeting with investors and you'll be brilliant there. And its clear that all reviewers are moved by such pure author's ambition. But if you believe that Machine Learning can be used for good in medicine, helping disabled and general embetterment of lives please try to look around little more.
And my favorite page 44.
Quoting: "The system allows a user to type their password into a web page—by waiting to see if the user logs in from eastern Europe and tries to get all their friends to install malware, you can get some evidence if the password was stolen or not." Yes, Geoff, you right, users from eastern Europe should not be allowed to use computers because your toaster sells can be harmed.
And my favorite page 44.
Quoting: "The system allows a user to type their password into a web page—by waiting to see if the user logs in from eastern Europe and tries to get all their friends to install malware, you can get some evidence if the password was stolen or not." Yes, Geoff, you right, users from eastern Europe should not be allowed to use computers because your toaster sells can be harmed.
September 18, 2019
A great overview of the implementation and orchestration of Intelligent Systems. Unlike most other books, the focus is on the process of using ML to provide value, rather than on the intricacies of any specific ML approach. Overall exactly what I was looking for.
Displaying 1 - 6 of 6 reviews