In recent years, a lot of attention has been attracted to sensors in mobile devices. Currently, the location of a user is the most commonly used kind of context information, but as more and more physical as well as virtual sensors are integrated into mobile devices and applications, new kinds of context-based services are conceivable. This in turn, leads to enhanced requirements for client-side applications and services, resulting in even more development effort and hence an increased time-to-market. Development frameworks for mobile devices try to counteract this trend by offering enhanced APIs, supporting the development of complex software components. A prominent example which has caused a lot of attention in recent years is the Android platform, which offers more higher-level programming concepts compared to other platforms for mobile devices, but unfortunately only primitive support for context mediation. In this paper, we therefore introduce a software component for easing the mediation of context data between mobile sensors and complex context-based applications and services. We analyze the requirements for mobile applications imposed by such complex services, present a generic high-level architecture of a software component meeting these requirements and highlight some implementation details of our prototypical implementation SenseDroid for the Android platform.