The PYTHON531-PYTHON534 Machine Learning Operations (MLOps) series provides an essential foundation for understanding the operations and practices that bridge machine learning development with reliable deployment and maintenance. Beginning with the background of current machine learning operations, we explore the challenges organizations face in scaling machine learning solutions, including issues in reproducibility, monitoring, and collaboration across teams. Central to this series of courses is the exploration of MLOps practices, covering the full machine learning lifecycle. We also outline the key roles involved in successful MLOps implementations, including data scientists, machine learning engineers, and data engineers. By the end of this module, learners will gain a comprehensive understanding of how MLOps enables scalable, reliable, and efficient machine learning operations within modern organizations.
- Prerequisite: Learners have completed PYTHON531 - Introduction to MLOps with MLFlow & PYTHON532 - Experiment Tracking with MLFlow
In PYTHON533 - Projects with MLFlow we introduce the MLflow Projects component, which provides a standardised way to package, organize, and share machine learning code for reproducible research and deployment. We start with an overview of MLflow Projects, covering its key concepts and explaining the structure and purpose of the MLproject file—a YAML configuration that defines dependencies, parameters, and entry points for executing machine learning workflows.
We will then move on to implement a basic MLflow project, gaining practical experience in defining a project structure and executing it via the MLflow CLI. Building on this foundation, we demonstrate how to include parameters in the MLproject file to allow flexible experiment configurations and how to add multiple entry points to support modular workflows with specific tasks, such as data preprocessing, training, and evaluation.
By the end of this course, learners will have a comprehensive understanding of MLflow Projects, enabling them to create, share, and execute machine learning code in a consistent and collaborative manner.