In this paper, we consider a multi-user massive MIMO network with hybrid beamforming architecture at the base station. The objective is to jointly perform user selection and design analog-digital hybrid beamformers in order to maximize a given utility function while satisfying various pertinent constraints. The problem is combinatorial and impractical to solve
optimally for a large system. In order to overcome the problem we develop a low-complexity heuristic algorithm. We propose a novel metric to judiciously select a predefined number of appropriate users and to pick congruent analog beamformers from a given dictionary based on the sparse regression techniques. Complementary digital beamformers are then designed for the selected users and fixed analog beamformers. Through simulations we analyze the performance of the proposed technique and compare it with the performance of standard techniques.