Introducing the CI/CD methodology
The CI/CD pattern has become a very popular methodology to automate the development and release of software into production. The main idea behind this practice is to make incremental, reliable, and frequent software code changes, and then deploy these changes automatically and seamlessly into production.
While this practice has been around for several years and employed by many DevOps engineers, the practice is starting to gain traction within the ML practitioner community, in the form of MLOps or Machine Learning Operations. However, before diving into how this methodology can be applied to ML, let's familiarize ourselves with the specific steps of the process, starting with CI.
Introducing the CI part of CI/CD
At a high level, the CI part of CI/CD comprises four key stages; Figure 4.1 shows a high-level overview of what these stages are:
Figure 4.1 – An overview of the stages in CI
As you can see from...
