MLOps and Cloud Engineering

Course description

This course introduces students to the principles and practices of deploying machine learning solutions in production environments. Students learn how to structure, automate, monitor, and manage ML systems at scale using modern MLOps tools and frameworks.

By working through deployment scenarios, students gain hands-on experience with containerization, cloud computing, model serving, and continuous integration/continuous deployment (CI/CD) pipelines. Emphasis is placed on reproducibility, scalability, and maintainability across local, on-premise, and cloud environments.

Course content

MLOps Concepts and Lifecycle
Reproducible Experiments and Tracking with MLflow and Azure ML
Continuous Integration / Deployment (CI/CD) with GitHub Actions
Deployment Options and Strategies (Real-Time vs Batch; Blue-Green, Canary, Rolling, A/B)
Containerization with Docker
Deployment Targets: Local, On-Premise (Portainer), and Cloud
Using Azure Machine Learning Services
Operational and Business Monitoring (Latency, Errors, Data & Concept Drift)
Continuous Retraining and Data Flywheels (User-Feedback Loops)
Version Control of Code, Data, and Models

Prerequisites

This course builds on earlier work in which students developed machine learning models for natural language processing (Y2A1) or computer vision (Y2B1). These models will now be prepared for production deployment.
Completion of all Year 1 courses in the Applied Data Science & AI programme.