Yes you can do it via the mlflow python API. However know that when you do that and move the experiment location to S3 the Databricks UI interface doesnt allow interaction with the artifacts etc. the mlflow python API is rich with the abilities to interact with the experiment metadata!