From Development to Deployment: An Approach to MLOps Monitoring for Machine Learning Model Operationalization
Anas Bodor, Meriem Hnida, Najima Daoudi
Abstract
Machine Learning Operations (MLOps) has emerged as an innovative and critical discipline in the development and deployment of Machine Learning models, ensuring smooth and efficient model management throughout their lifecycle. In this paper, we propose an implementation of MLOps concept using a pipeline. For instance, we propose a solution to automate the deployment of Machine Learning Models in production, while ensuring the monitoring of various predefined metrics, in order to guarantee the prediction quality of a model throughout its deployment. Maintaining the ML model in production, while exploring the possibility of using the AutoML concept, is essential to guarantee simplified, automated model management, enabling effective adaptation to change in data and continuous improvement in system performance.