Work in Science Studies and Computer Supported Cooperative Work often examines teams of natural scientists and computer scientists collaborating to design and build infrastructure from the top down. In contrast, this talk examines the collaborative work of a radio astronomy group at the University of Washington crafting a data processing and analysis infrastructure that has grown to support researchers across an international empirical cosmology consortium. This qualitative study examines the co-production of software and data products, which enables us to interrogate and reflect upon the deeply intertwined nature of these two types of research products, among others. Our work inspects the practices that support the emergence and growth of new infrastructures within a larger constellation of scientific knowledge and experiment designs, and those experiment’s resulting hardware infrastructures. Using the conceptualizations of coordinative artifacts, this talk will discuss how a group of radio astronomers crafts their software and data products as they pursue their research questions in a rigorous and replicable manner. We examine the group’s adoption of publicly available coordinative systems, in particular GitHub and Dropbox, as part of emergent software and data practices that comprise their overall data-intensive research practices. Finally, we assess implications for the sharing of research products among cosmology researchers and beyond.