SEAD – a project funded by the US National Science Foundation’s DataNet program – has spent the last five years designing, building, and deploying an integrated set of services to better connect scientists’ research workflows to data publication and preservation activities. Throughout the project, SEAD has promoted the concept and practice of “active curation,” which consists of capturing data and metadata early and refining it throughout the data life cycle. In promoting active curation, our team saw an opportunity to develop tools that would help scientists better manage data for their own use, improve team coordination around data, implement practices that would serve the data better over time, and seamlessly connect with data repositories to ease the burden of sharing and publishing. SEAD has worked with 30 projects, dozens of researchers, and hundreds of thousands of files, providing us with ample opportunities to learn about data and metadata, integrating with researchers’ workflows, and building tools and services for data. In this paper, we discuss the lessons we have learned and suggest how this might guide future data infrastructure development efforts.