Sure, as the the underlying Java/Scala code that interprets the microflow may be multi-threaded. Either because it is explicitly so or because it calls functions that have multi-threaded implementations. But do not expect the microflow interpreter to recognize paths that can be executed in parallel and execute those in different threads (it would be nice if it does this but I would be surprised if it is the case, also, in practice this would usually provide little gain as everything involving object state needs to be synchrnoised anyway).