What do you think?
Rate this book
Machine Learning Design Patterns: Solutions to Common Challenges in Data Preparation, Model Building, and MLOps
The design patterns in this book capture best practices and solutions to recurring problems in machine learning. Authors Valliappa Lakshmanan, Sara Robinson, and Michael Munn catalog the first tried-and-proven methods to help engineers tackle problems that frequently crop up during the ML process. These design patterns codify the experience of hundreds of experts into advice you can easily follow.
The authors, three Google Cloud engineers, describe 30 patterns for data and problem representation, operationalization, repeatability, reproducibility, flexibility, explainability, and fairness. Each pattern includes a description of the problem, a variety of potential solutions, and recommendations for choosing the most appropriate remedy for your situation.
You’ll learn how to:
Identify and mitigate common challenges when training, evaluating, and deploying ML models
Represent data for different ML model types, including embeddings, feature crosses, and more
Choose the right model type for specific problems
Build a robust training loop that uses checkpoints, distribution strategy, and hyperparameter tuning
Deploy scalable ML systems that you can retrain and update to reflect new data
Interpret model predictions for stakeholders and ensure that models are treating users fairly
The authors, three Google Cloud engineers, describe 30 patterns for data and problem representation, operationalization, repeatability, reproducibility, flexibility, explainability, and fairness. Each pattern includes a description of the problem, a variety of potential solutions, and recommendations for choosing the most appropriate remedy for your situation.
You’ll learn how to:
Identify and mitigate common challenges when training, evaluating, and deploying ML models
Represent data for different ML model types, including embeddings, feature crosses, and more
Choose the right model type for specific problems
Build a robust training loop that uses checkpoints, distribution strategy, and hyperparameter tuning
Deploy scalable ML systems that you can retrain and update to reflect new data
Interpret model predictions for stakeholders and ensure that models are treating users fairly
400 pages, Paperback
Published November 3, 2020
202 people are currently reading
1,012 people want to read
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 26 of 26 reviews
February 6, 2023
This was a fantastic book that builds a common vocabulary and patterns for understanding and dealing with various machine learning problems. Won't even take off a star for vendor-specific implementations. I got so much out of it. Notes here for anyone who wants them.
February 1, 2021
This is a must-have book for every data scientist and machine learning engineer. While most of the ideas are not new, this is a very good catalog and a useful reference when solving an ML problem.
January 2, 2022
Very useful book for anyone looking for best practices, especially when it comes to anyone moving from data science to machine learning engineering job or MLOps. Don’t expect explanations for algorithms here, but mainly how to deploy your model, monitor them, plan their training, etc.
For me, lots of the information are not new to me, guess this will be the case for anyone who has been in the same field for a while already, however I find new stuff to learn every chapter still.
I have seen criticism for the book for being too Google-centric. I can see that, with all the examples based on Big Query, TensorFlow, etc. But that did not bother me at all despite the fact that I haven’t used Google Cloud technology before. For me it was easy to generalise and see the concepts and ignore the specifics of Google’s products and their implementation.
For me, lots of the information are not new to me, guess this will be the case for anyone who has been in the same field for a while already, however I find new stuff to learn every chapter still.
I have seen criticism for the book for being too Google-centric. I can see that, with all the examples based on Big Query, TensorFlow, etc. But that did not bother me at all despite the fact that I haven’t used Google Cloud technology before. For me it was easy to generalise and see the concepts and ignore the specifics of Google’s products and their implementation.
January 27, 2022
More like a cookbook of recipes. Too focused on promoting GCP/Tensorflow/Big Query.
August 7, 2022
Great. Covers a pretty broad spectrum of ML problems and from a practical standpoint. If you already have experience in ML, or good foundational knowledge, you will probably want to skip some of the chapters. On the other hand, I think this can be a great resource for those who skipped the fundamentals. Unfortunately, a lot of the hands-on content is geared towards Google's ML infra. (can be a benefit if that's what you intend to use). I found the chapters in responsible AI the most interesting.
December 31, 2021
Great reference.
July 28, 2022
Good advice and structured in a very helpful way; main complaint is that I couldn't get any of the code examples to work at all despite trying three different methods across two devices
March 31, 2021
I actually read a technical book from the beginning to the end. Rated here so I can get similar recommendations.
January 24, 2021
The process of reading this book was not only informative and insightful, but also enjoying. Authors clearly have a good idea in broadly addressing ML Lifecycle with different Design Patterns that could make to reach end-goal (i.e. effective model performance) in machine learning as swiftly and accurately as possible with future development. They discussed ins and outs of every ML problems and how to tackle them with various methods, moreover, they tend to discuss ML algorithms not only in terms of Deep Learning or any other specific algorithm, but in general.
The book is advised to use as a guideline for future ML projects, because it gives a good explanation to reach maximum performance for ML models. The last chapter clearly and descriptively summarizes the whole book so to use like a cheat-sheet.
The book is advised to use as a guideline for future ML projects, because it gives a good explanation to reach maximum performance for ML models. The last chapter clearly and descriptively summarizes the whole book so to use like a cheat-sheet.
May 7, 2022
It’s a nice and cohesive overview of ML design techniques. However, the book is too focused on the Google Cloud and Tensorflow way of doing MLOps and Machine Learning in general
April 28, 2024
I feel a little conflicted giving this book a lower rating, and though I'm sure I'll reach for it again someday as a reference, I think it's still a pretty fair assessment of its strengths and weaknesses.
The good:
The subjects covered encompass the entire ML lifecycle, from data ingestion and cleaning, to model tuning and deployment. If you have some experience in this area, a lot of these techniques will be familiar to you, but even then, the way the authors frame each of them as a "pattern" is exquisitely done.
The book is broken down into main themes, and then further into specific patterns. For each of these they present the problem, solution, why it works, and tradeoffs/alternatives. This organizational style makes it very valuable as a reference, and as a refresher for why we do things sometimes.
The not so good:
The book should have been named "Machine Learning Design Patterns...WITH GOOGLE CLOUD SERVICES, LOL".
Seriously, while the authors attempt to provide the reader with some variety regarding the tools one might use to enforce these aforementioned patterns, the examples invariably end up coming back to the Google Cloud Stack. This means you should expect lots of TensorFlow examples, BigQuery ML, etc. Which is fine, but then don't say you'll present "options" when 99% are just one. Anyways...
They really can't help themselves, and sometimes hilarity ensues. Case-and-point: there's a chapter where they attempt to implement one of their patterns using XGBoost, saying something along the lines of "the last few examples have been done using TensorFlow, so will take a look at other tools..." They make the smallest effort to get it going but then they construct this weird hypothetical case so they can get back to using what they know and like, and it goes like this, "but now suppose someone comes in and wants to use TENSORFLOW mkay? Let's do it..."
So there you have it. I think it is still worth to keep as a reference and to fill in the gaps and blind spots you'll ultimately develop as an ML practitioner, but if you are expecting something more hands-on with non-Google stack tools, look elsewhere.
Recommended with caveats as above.
The good:
The subjects covered encompass the entire ML lifecycle, from data ingestion and cleaning, to model tuning and deployment. If you have some experience in this area, a lot of these techniques will be familiar to you, but even then, the way the authors frame each of them as a "pattern" is exquisitely done.
The book is broken down into main themes, and then further into specific patterns. For each of these they present the problem, solution, why it works, and tradeoffs/alternatives. This organizational style makes it very valuable as a reference, and as a refresher for why we do things sometimes.
The not so good:
The book should have been named "Machine Learning Design Patterns...WITH GOOGLE CLOUD SERVICES, LOL".
Seriously, while the authors attempt to provide the reader with some variety regarding the tools one might use to enforce these aforementioned patterns, the examples invariably end up coming back to the Google Cloud Stack. This means you should expect lots of TensorFlow examples, BigQuery ML, etc. Which is fine, but then don't say you'll present "options" when 99% are just one. Anyways...
They really can't help themselves, and sometimes hilarity ensues. Case-and-point: there's a chapter where they attempt to implement one of their patterns using XGBoost, saying something along the lines of "the last few examples have been done using TensorFlow, so will take a look at other tools..." They make the smallest effort to get it going but then they construct this weird hypothetical case so they can get back to using what they know and like, and it goes like this, "but now suppose someone comes in and wants to use TENSORFLOW mkay? Let's do it..."
So there you have it. I think it is still worth to keep as a reference and to fill in the gaps and blind spots you'll ultimately develop as an ML practitioner, but if you are expecting something more hands-on with non-Google stack tools, look elsewhere.
Recommended with caveats as above.
April 17, 2022
o que chamou atenção nesse livro:
* banho de loja de google cloud e bigquery, com, pasmem, até treinamento de modelo dentro do bigquery
* organização entre problema/solução/why it works/trade-offs forçada, quase sempre eu nem sabia em qual sub-seção eu tava
* explicações mal-feitas ou mal-justificadas e com carência de fontes externas
* tudo muito introdutório ou superficial, não consegui achar nada relevante ao meu caso de uso que eu já não tivesse ou aplicado antes ou pensado antes
* muito centrado em deep learning, e em alguns momentos não deixando isso explícito
* comentário bem desrespeitoso a analistas de dados no começo do livro
2 estrelas pq deve ser útil pra noobs e infelizmente ainda não tem abundância de alternativas de livros sobre o tema
* banho de loja de google cloud e bigquery, com, pasmem, até treinamento de modelo dentro do bigquery
* organização entre problema/solução/why it works/trade-offs forçada, quase sempre eu nem sabia em qual sub-seção eu tava
* explicações mal-feitas ou mal-justificadas e com carência de fontes externas
* tudo muito introdutório ou superficial, não consegui achar nada relevante ao meu caso de uso que eu já não tivesse ou aplicado antes ou pensado antes
* muito centrado em deep learning, e em alguns momentos não deixando isso explícito
* comentário bem desrespeitoso a analistas de dados no começo do livro
2 estrelas pq deve ser útil pra noobs e infelizmente ainda não tem abundância de alternativas de livros sobre o tema
April 21, 2021
A must read for intermediate ML practitioners who want to follow best practises around model building and deployment - minimising the latency and optimising the throughput. Being in this field for almost 7 years, I found most of the patterns relatable and structured. Really enjoyed reading the design patterns on resilient serving and reproducibility.
This book is not meant for those who are seeking to learn the basics of ML algorithms, however you will find a section on reframing the real world problem statements into practise using various ML algos.
Would recommend reading it at least once!
This book is not meant for those who are seeking to learn the basics of ML algorithms, however you will find a section on reframing the real world problem statements into practise using various ML algos.
Would recommend reading it at least once!
July 20, 2021
I totally disagree with the name of "Design Patterns", instead I would suggest something like "some best practices, real-world use cases, and a few design patterns in its infancy". However, the book is really good for people with experience implementing IA. Hopefully, in the future, we can refer to this book as the pioneer in the field.
August 21, 2022
One of the best books on Machine Learning systems. It covers many design patterns including both ML best practices and engineer best practices.
It also explains many concepts in ML such as Embedding, Ensemble, Transfer Learning, etc.
Overall, machine learning should be reproducible (pipeline), explainable, no biases.
It also explains many concepts in ML such as Embedding, Ensemble, Transfer Learning, etc.
Overall, machine learning should be reproducible (pipeline), explainable, no biases.
August 20, 2021
Great book for intermediate/advanced ML practitioners. It gave me many ideas on how to better solve problems with ML.
One note about polish translation - it's abysmal. It was easier for me to read english version from laptop screen than the paper one in polish I possess.
One note about polish translation - it's abysmal. It was easier for me to read english version from laptop screen than the paper one in polish I possess.
March 19, 2021
Insightful, clear example with reason and alternatives to solve problems
October 13, 2022
A book with many helpful design patterns. Definitely will come back to check it out in detail again to avoid some common pitfalls and find the best practice.
November 21, 2022
Very useful guide, covering the end-to-end process of ML engineering. I do wish the book could make use of more real industry examples, though.
December 11, 2022
A very practical book for data scientists and machine learning engineers, which helps solve most problems in building end-to-end ML models and pipelines.
January 6, 2023
Too much talk for promoting google products: GCP, Tensorflow, ...
April 6, 2023
It would be better as some online course or set of blog posts.
May 18, 2023
Good lil handy book for reference at work. I am going to recommend this book to my colleagues.
January 22, 2022
Very good catalog of machine learning terminology and practices. A must have to quickly check the definition, benefits and drawbacks of a practice through the index.
Exemples are often given in the context of Google Cloud, Keras and Tensorflow , however the concepts covered are applicable to other types of providers.
Also covers various parts of the lifecycle from experiment to validation to production.
Exemples are often given in the context of Google Cloud, Keras and Tensorflow , however the concepts covered are applicable to other types of providers.
Also covers various parts of the lifecycle from experiment to validation to production.
September 12, 2023
This is a good book overall, but too much focused on using Google cloud products. It almost feels like a marketing book by Google which they are selling and not even distributing free. I would have liked more focus on theory and concepts rather than code that too only in BigQuery, screenshots of Google cloud suite etc.
August 28, 2022
Really great collection, slightly vendor locked but in general applicable everywhere.
Displaying 1 - 26 of 26 reviews
Can't find what you're looking for?
Get help and learn more about the design.