This thesis investigates the development of research interests in the field of computer sci- ence by examining how researchers transition between research objects and how different collaborative relationships impact these transitions. The study begins with an exploration of the dual spheres of research activities: the idea sphere and the engineering sphere, along with the two primary epistemic faculties—intuition and agency—that enable researchers to navigate between them. Drawing on qualitative data from interviews and case studies, the thesis highlights how these faculties contribute to the generation and evolution of research interests. A case study of a computer scientist’s publication trajectory illustrates the dif- ference between local and non-local research interest transitions, showing how researcher’s shifts in research objects, accompanied by the reconstruction of epistemic faculties, mark significant changes in research interest. The quantitative analysis further investigate the relationship between collaborative rela- tionships and local and non-local interest transition. The study uses bibliometric data from the DBLP-D3 dataset and employs measures such as the Distance of Object Engagement (DOE) and the Distance of Community Affiliation (DCA). The findings reveal that while one-time collaborations contribute to incremental changes in research interest, recurrent col- laborations play a pivotal role in facilitating non-local transitions that require substantial knowledge and skill adjustments. This suggests that stable, intensive collaborators provide essential support, allowing researchers to adapt their intuition and agency when exploring new, unfamiliar research domains.