This paper introduces FLAMI, a software reference architecture for building big data-sharing repositories that adhere to the FAIR (Findability, Accessibility, Interoperability, and Reusability) principles. Inspired by the data lakehouse concept, FLAMI can extract, load, and transform large volumes of data and metadata from heterogeneous data providers into a unified storage solution. It integrates with existing infrastructures of external repositories, allowing data stored outside the lakehouse to be fetched and processed within its infrastructure. FLAMI also introduces a set of guidelines for its implementation. We validate our proposal through a case study that instantiates FLAMI to the context of a real-world seismology dataset, exploiting analytical queries that encompass the needs of geoscientists. We also employ a framework to assess FLAMI’s FAIR compliance, achieving 100% in its conceptual version and over 73% in the seismology instantiation.