Almost every ML system you will operate runs on Linux, and almost every change to it should go through version control. This module builds the habits that make machine learning work reproducible: knowing your way around the terminal, tracking every change to your code, and doing the same for the data and models that code depends on.
You'll start with a quick look at Linux and what makes it the default environment for servers and ML infrastructure. Then you'll learn how Git tracks changes through commits and branches. Finally, you'll see why Git alone isn't enough for ML projects, since datasets and model files are too large to commit, and how DVC (Data Version Control) fills that gap by versioning data and models alongside your code.
With code, data, and models all under version control, any experiment can be reproduced later. That sets you up for the next module, Data Pipelines, where you'll start moving that data through automated workflows.
A fast introduction to the Linux operating system, its history, and why it powers most servers and cloud infrastructure.
The core ideas behind Git: repositories, staging files, commits as snapshots, and branching and merging.
A code-free overview of why data science projects need version control for data and models, and where DVC fits in.
A beginner-friendly walkthrough of using DVC together with Git to version datasets, models, and more.