The availability of affordable high-performance computing, large data archives, advanced visualization capabilities, together with distributed workflow management through grid middleware, has made sophisticated computer simulations an indispensable part of cutting-edge research. This is particularly true for interdisciplinary questions, coupling domain knowledge from disciplines previously studied separately. Computer science (CS) has always been at the core of providing the software infrastructure of such endeavors, but typically in a fairly generic fashion. On the other hand, there is ample scope for injecting CS know-how directly into numerical simulations. One such example is Automatic Differentiation (AD), a technique for augmenting computer programs with statements for the computation of derivatives based on the chain rule of differential calculus. This talk will report on infrastructural, educational and research activities in computational science and engineering currently under way at RWTH Aachen University, illustrate the role of CS in that process, and showcase some recent developments in AD.