In this work, we present a methodology for the prediction of bulk properties of granular materials by employing a platform developed by the authors called NeuroDEM®, which combines artificial neural networks (ANNs) and the discrete element method (DEM). As a model problem, we are first interested in predicting the effective (i.e., bulk) thermal conductivity of a granular assembly. Accordingly, an ANN is trained with the help of computed effective conductivities of various different assemblies, obtained through several simulations with our DEM code. Heat transfer by convection and radiation are not considered as to isolate the conduction contribution and allow for a better estimate of the assembly’s effective response. Then, we apply the developed methodology to the context of additive manufacturing. Here, also as a model problem, we are interested in predicting the final bulk porosity of a manufactured piece from a powder bed by simulating a selective laser sintering (SLS) process. Similarly to the first problem, we train our ANN with the help of several simulations, which provide the bulk porosity for each case. To minimize the ANN weights, we use different optimizers (gradient descent, genetic algorithms and the combination of both) to assess which leads to a better performance. From the obtained results, we conclude that the proposed methodology enables an efficient tool for the prediction of bulk properties of granular assemblies over a wide range of values in different applications.