In a collaborative machine learning project, you need to ensure that all team members have access to the same Python library versions across different notebooks in the Databricks Runtime for Machine Learning. What is the best practice to achieve this?