New machine learning algorithms are often first scripted at university labs, gluing together several languages such as shell scripting, Python, R, MATLAB, Scala, or C++ to provide a new concept and theoretically analyze its properties. An algorithm might take a long path of refactoring before it lands in a library with standardized input or output and interfaces. While Python, R, and MATLAB are quite popular, they are mainly used for scripting, research, and experimenting. Java, on the other hand, is the de facto enterprise language, which could be attributed to static typing, robust IDE support, good maintainability, as well as decent threading model and high performance concurrent data structure libraries. Moreover, there are already many Java libraries available for machine learning, which makes it really convenient to apply them in existing Java applications and leverage powerful machine learning capabilities.