Hi,
The Data Science Virtual Machine is designed for data exploration, model development, model validation, and related activities.
By "production environment", I assume you mean something like creating a REST API from a model, or large-scale batch scoring, or using the trained model in some other capacity. The DSVM is not designed for these types of production uses and is
not a hardened environment. We recommend that you use
Azure Machine Learning Studio,
Azure Machine Learning Services,
Machine Learning Server, or another product that is specifically designed for production use. These solutions will provide the security, scale, and support needs you are looking for.