Machine Learning Operations
Objectives
Being MLOps a practice that focuses on deploying, testing, monitoring, automating ML models in large-scale production environments, and recognizing its importance on the quality standards and simplification of data management process, the students will learn the best practices of its implementation. As an answer to the problem of scaling ML operations to the needs of the business and/or users of ML models in a standard way, the students will understand how, after the implementation of the MLOps, It?s easier to align models with business needs, as well as regulatory requirements.
General characterization
Code
200293
Credits
3.5
Responsible teacher
Nuno Filipe Rosa Agostinho
Hours
Weekly - Available soon
Total - Available soon
Teaching language
Portuguese. If there are Erasmus students, classes will be taught in English
Prerequisites
TBD
Bibliography
Introducing MLOps: How to Scale Machine Learning in the Enterprise 1st Edition by Mark Treveil et al.
Machine Learning Engineering by Andriy Burkov
Beginning MLOps with MLFlow: Deploy Models in AWS SageMaker, Google Cloud, and Microsoft Azure by Sridhar Alla and Suman Kalyan Adari
Teaching method
TBD
Evaluation method
Final project of creating a ML end-to-end cycle 70%
Exam 30%
Subject matter
1-MLOps Fundamentals
2-Model experimentation, deployment, and testing using MLFlow
3-Model monitoring: data drift vs concept drift
4-Model monitoring with NannyML
5-How to make a ML project development much more efficient using Kedro
6-Understanding the Main Kubernetes
7-Introducing Kubeflow pipelines
Programs
Programs where the course is taught: