As the complexity of scientific software increases, scientists are often expected to be experts in their own domain as well as in the construction and maintenance of their software. To support this paradigm, numerous specialized approaches have emerged. Leveraging the underlying expertise and motivational factors that drive scientific software development practices, the authors are developing a hybrid scientific software process improvement framework (SciSPIF) to allow scientific software developers to "self-drive"' their SE process improvement activities according to their own goals. Here, the authors report on the strategy for designing and building SciSPIF and provide preliminary insights.