Engineering MLOps

• Credit value: 30 credits at Level 6
• Convenor: Dr Stelios Sotiriadis
• Assessment: a mini-project (40%) and a two-hour open book examination (60%)

The Engineering MLOps module enables you to develop the skills required to deploy, monitor and maintain machine learning models in production environments. You will learn best practices in continuous integration and deployment (CI/CD), model versioning and performance monitoring.

This module prepares you for a critical role such as an MLOps engineer, DevOps engineer or cloud engineer in ensuring robust and scalable AI solutions in different verticals.

Indicative syllabus

• Introduction to MLOps: an overview of MLOps concepts, principles and the importance of operationalising machine learning models
• Continuous integration and continuous deployment (CI/CD): best practices for automating the deployment of machine learning models
• Model versioning: techniques for managing different versions of machine learning models
• Monitoring and logging: tools and methods for monitoring model performance and logging data for analysis
• Scalability and reliability: ensuring models are scalable and reliable in production environments
• Infrastructure as code (IaC): using IaC tools to manage and provision infrastructure for machine learning models
• Data management: handling data pipelines, data versioning and ensuring data quality
• Security and compliance: addressing security concerns and ensuring compliance with regulations in MLOps practices
• Collaboration and communication: working effectively with cross-functional teams, including data scientists, engineers and stakeholders
• Case studies and real-world applications: practical examples and hands-on projects to apply MLOps concepts in real-world scenarios

By the end of this module you will be able to:

• demonstrate how the core principles and practices of MLOps can be applied
• apply continuous integration and continuous deployment (CI/CD) techniques to machine learning workflows
• evaluate the performance and reliability of machine learning models in production environments
• analyse the challenges and solutions in operationalising machine learning models
• critically evaluate different MLOps tools and frameworks for their effectiveness in various scenarios
• formulate strategies for managing model versioning and data pipelines
• synthesise knowledge from machine learning, software engineering and DevOps to optimise AI workflows
• implement CI/CD pipelines for machine learning models using industry-standard tools
• use monitoring and logging tools to track model performance and detect issues
• develop scalable and reliable infrastructure for deploying machine learning models
• conduct experiments to improve model performance and operational efficiency
• communicate technical concepts and findings effectively to both technical and non-technical stakeholders
• work independently and collaboratively on MLOps projects, demonstrating strong teamwork and project management skills
• develop self-awareness and continuous learning skills by staying updated with the latest advancements in MLOps
• demonstrate an increased awareness of ethical practices and the social impact of operationalising AI technologies.