You have a custom job that runs on Vertex AI on a weekly basis. This job is part of a proprietary ML workflow that generates datasets, models, and custom artifacts, and stores them in a Cloud Storage bucket. Over time, many versions of datasets and models are created. Due to compliance requirements, your company must be able to track which model was used to make a specific prediction and have access to all related artifacts for each model. How should you configure your workflows to meet these compliance and tracking requirements?