MLOps Fundamentals: CI/CD/CT Pipelines of ML with Azure Demo

Important Note: The goal of this course is to teach MLOps foundations rather than Azure Machine Learning. The Azure demo part is presented as proof of concept to demonstrate how an end-to-end MLOps project works. All of the pipeline’s codes are, however, described.

For a long time, data scientists have been experimenting with machine learning models, but in order to give meaningful commercial value, the models must be operationalized, or put into production. Unfortunately, due to present limitations and a lack of systemization in the machine learning lifecycle, 80 percent of models never make it to production and remain a research project.

According to market buzz, 2021 will be the year of MLOps, with MLOps becoming the required skill set for Enterprise ML initiatives.

What exactly is covered in the course?

• MLOps principles and essential concepts.
• What were the difficulties in managing the typical machine learning lifecycle?
• How MLOps is solving these difficulties while increasing ML process flexibility and automation.
• MLOps is founded on a set of standards and concepts.
• The MLOps pipelines for continuous integration (CI), continuous delivery (CD), and continuous training (CT).
• MLOps come in a variety of maturity levels.
• Comparisons of MLOps tools and MLOps platforms.
• A crash lesson in Azure Machine Learning components in a few minutes.
• An Azure DevOps and Azure Machine Learning-based end-to-end CI/CD MLOps pipeline for a case study.