Practical MLOps

Chapter 2. MLOps Foundations

By Noah Gift

Medical school was a woeful experience, an endless litany of fact, whose origins were rarely explained and whose usefulness was infrequently justified. My distaste for rote learning and my questioning attitude were not shared by most of the class of 96 students. This was particularly evident on one occasion when a biochemistry lecturer claimed to be deriving the Nernst equation. The class was faithfully copying what he wrote on the board. Having only a year before taken the Pchem course for chemistry majors at UCLA, I thought he was bluffing.

“Where did you get that value for k?” I asked.

The class shouted me down: “Let him finish! Just copy it.”

Dr. Joseph Bogen

Having a solid foundation to build on is critical to any technical endeavor. In this chapter, several key building blocks set the foundation for the rest of the book. When dealing with students new to data science and machine learning, I have commonly encountered misconceptions about the items covered in this chapter. This chapter aims to build a strong foundation for using MLOps methodologies.

Bash and the Linux Command Line

Most machine learning happens in the cloud, and most cloud platforms assume you will interact with it to some degree with the terminal. As such, it is critical to know the basics of the Linux command line to do MLOps. This section aims to bootstrap you with just enough knowledge to ensure you have success doing MLOps.

There is often a look of both shock ...