Docker is a tool where one reads the documentation and be able to use it. ML, much like math, is a discipline. Sure you can read textbooks and the latest papers. But deep understanding and experience are what make it useful. Sadly these will take much longer.