Research of the ci/cd approach adaptation possibilities to the development of machine learning models

In this paper, we proposed a method for the automated deployment of machine learning algorithms based on the Splunk Enterprise software product and the Splunk Machine Learning Toolkit application for IT. The implementation of this method will make it possible to deploy ML systems in the shortest possible time, make changes to their structural units with minimal impact on other components and adapt ML models to changes in input data, transfer the system to another environment or cloud service provider. The advantage of using this method is the ability to monitor user activity and, if necessary, detect anomalies in their behavior. Anomalies are detected among system/audit log data. After installing the required data sources on the servers for monitoring, the data will be received on the indexer and will be available for further processing and training of the ML model.