Management
Management of any system involves security, safety, monitoring and performance, and optimization. In the case of SQL Server Machine Learning Services, the safety portion (backups, availability, and the like) are part of the database environment. Performance tuning involves optimizing the T-SQL and language-specific code and calls. That leaves you with a specific set of processes and tools for security, as well as monitoring and performance.
Security
For the most part, the security for using Machine Learning Services follows the same model as other SQL Server securables. The person or SQL Server principal calling the Machine Learning Services extensibility framework functions needs to be a Windows or SQL Server database user, must have access to the tables or views they are passing in, the ability to write data out (if they do that with the returned data), and be able to create stored procedures if they are making new code to run the models.
There are differences...