The production and use of software pipelines is a key component of much modern scientific research. We present emerging findings from our qualitative, social science study of a radio astronomy group developing software pipelines as they produce a data processing infrastructure. This paper examines how these researchers co-produce data products and software pipelines to enact their research infrastructure. We investigate the work of co-producing data and software to illustrate that to better support data-intensive science, we need to understand the practices that enable and produce data products, software, and ultimately infrastructures.