In this example embodiment, where the monolithic application comprises an example trading application, microservice generation system 102 can train model component 108 to achieve the objective described above to minimize the negative log likelihood of observing these neighborhoods of node conditional on the node embedding by employing equation (1) defined below.