This presentation will report on the development of a cutting-edge "Next Generation Data Repository" (NGDR), which is currently being developed by Antleaf and Cottage Labs for the National Institute of Materials Science (Japan), and which is scheduled to be launched at the end of March 2019. The NGDR has been developed to adopt several of the "behaviours" defined by COAR's Next Generation Repositories Working Group and, as such, implements technologies such as ResourceSync. We will also focus on the complex metadata requirements for a data repository in the domain of materials science, and will describe the approaches to managing and exploiting sophisticated metadata from a variety of sources including automated generation from laboratory equipment, crowd-sourced vocabularies, and standard SKOS-based controlled vocabularies. Finally, the challenge of integrating the repository with a number of other systems in the research 'pipeline' will be described.