Using examples from our recent work, I will present two approaches used by our group to develop porous materials to be applied in a variety of separation applications. The first one is a curiosity-driven approach, which we employ either for new separation challenges for which benchmark and/or performance targets are ill-defined or in the context of emerging adsorbent materials. I will illustrate this aspect through our recent findings on porous boron nitride. Our second approach, based on molecular engineering, accelerates materials development for well-defined applications. Such approach links molecular modelling, experiments and process modelling to quickly identify the best adsorbent(s) for a given separation. In this context, I will present our work on the use of metal organic frameworks for industrially relevant gas separations.