Abstract
This paper investigates the development of domain-specific metadata schemata to support FAIR data management within specialised research communities. We report on two case studies, the Core Metadata Schema for Learner Corpora (LC-meta) and the metadata schema developed for interpreting corpora within the Unified Interpreting Corpus (UNIC) project. We address how metadata are described and standardised in these projects, the principles that informed the schema design, and the infrastructural challenges that arise when semantically rich, community-specific metadata are to be represented within generic discovery frameworks. Our findings underline the importance of domain-oriented approaches to metadata standardisation, iterative design and stakeholder involvement, and suggest a more active role for CLARIN in supporting long-term sustainability, technical mediation, and shared stewardship.