I would consider defining MLOps more broadly than the article (which only discusses production). For example, it can be as simple as wondering how to organize,

I would consider defining MLOps more broadly than the article (which only discusses production). For example, it can be as simple as wondering how to organize, collaborate on, and versionize lots of Jupyter notebooks! Or how to keep track of and compare different models/hyperparameters/datasets, not only individually but also across a team. Both are not discussed in typical data science courses.