If you’re looking to take your machine learning projects to the next level, MLOps is an essential part of the process. In this article, we’ll provide you with a practical tutorial on how to structure your projects for MLOps, using the classic handwritten digit classification problem as an example. We’ll take you step-by-step through the process of creating a basic project template that you can use to organize your own projects. By the end of this tutorial, you’ll have a solid understanding of MLOps principles and how to apply them to your own projects. However, if you’re new to MLOps, we recommend starting with my beginner-friendly tutorial to get up to speed. So let’s dive in and take your ML projects to the next level!
Table of contents:
· 1. Introduction
· 2. MLOps
??? 2.1. Business problem
??? 2.2. Data engineering
??? 2.3. Machine learning model engineering
??? 2.4. Code engineering
· 3. Project structure
??? 3.1. Cookiecutter Data Science
· 4. MLOps project structure
??? 4.1. Starting a new MLOps project
??? 4.2. Using MLOps project template for handwritten digits classification
??? 4.3. How to run your project?
· 5. Conclusion
My MLOps tutorials:
- Tutorial 1: A Beginner-Friendly Introduction to MLOps
- Tutorial 2: Structuring Your Machine Learning Project with MLOps in Mind
[I will be updating this list as I publish articles on the subject]
1. Introduction
In the previous tutorial, we defined MLOps as a way to design, build, and deploy machine learning models in an efficient, optimized, and organized manner. This is achieved by combining a set of techniques, practices, and tools that are often discussed within the context of the MLOps lifecycle.
In the MLOps lifecycle, the first step after understanding the problem is to structure your project. This is typically done by using a template, whether it’s a company template, a public template, or your own template, as we will see later in this tutorial.