Going from academia to industry is daunting.
The inconvenient truth is that courses do not go beyond model training.
The solution: These books close the knowledge gap for deployment and MLOps.
Designing Machine Learning Systems
by Chip Huyen
The most highly anticipated MLOps book!
This book is for people that already know ML and possibly deployed models.Learn to design reproducible data systems and ML deployment in depth.
Pick up the book here.
Also, this is probably the cleanest Github repo out of the bunch:
Building Machine Learning Powered Applications: Going from Idea to Product
This book serves as a lightweight intro to concepts behind powering apps with ML.
The main example walks through building an ML-assisted writing app.
Get a copy of the book here.
Machine Learning Design Patterns
by Lak Lakshmanan
This book starts with basic patterns around data, features, and labels.
The book goes into detail on deploying Tensorflow models while keeping reproducibility, responsibility and resilience in mind.
Available for purchase here.
There’s also an open-sourced repo with code examples for the book.
Building Machine Learning Pipelines
by Hannes Hapke & Catherine Nelson
How do you automate and scale machine learning?
First, this book goes into detail on TFX and TF Serving.
Then to automate and scale goes into Airflow and Kubeflow.
Get scaling here and check out the code repo here.
In this short post, we looked at 4 amazing books for MLOps.
- Designing Machine Learning Systems
- Building Machine Learning Powered Applications
- Machine Learning Design Patterns
- Building Machine Learning Pipelines
Check them out for a thorough education!