The performance optimization of a geothermal reservoir depends on the ability of combining available data with the deep understanding of mutually coupled processes over the whole hierarchy of scales, from the pores to the reservoir. These include hydrothermal flows in heterogeneous porous media, mechanical processes, and reactive transport, among others. This work focuses on the geophysical, mathematical, and computational challenges related to the modeling and simulation of geothermal reservoirs, and preliminary results concerning the optimal placement of multi-well systems for geothermal energy production. In this talk, we will overview the results of the ongoing collaboration between LIAG and WIAS, started in the framework of the last MMS days and supported by a MMS Seed Money Grant in 2018. The talk will be jointly presented by A. Caiazzo (WIAS) and E. Meneses (LIAG)