DevOps practices have increasingly been applied to software development and engineering, as well as the machine learning lifecycle – in a process also known as MLOps. Today, many companies and professionals have been working and writing on this topic. However, in the academic and scientific literature, few results can be found on MLOps and how to implement it efficiently. This paper presents five essential steps to guide the understanding and practice of MLOps, which, based on the authors’ research and experience, can assist in its effective implementation. The study aims to serve as a reference guide for all those who wish to learn more about the topic and intend to implement MLOps practices in the development of their systems.