Successful architecting of complex systems requires reconciling heterogeneous viewpoints expressed by the stakeholders involved in the development process, including domain and technical experts, users and managers. Most of the time, each concern is analyzed by experts using well-fitted specific tools to produce their point of view on a solution. This results in a set of models with various technical spaces, formalisms and paradigms. Ensuring global consistency, maintaining traceability and building cross- concerns views in that context is challenging. In order to address this issues, we initiated the development of a tooling that provides support for building conceptual views expanding upon existing models and tools. It has been applied to uses cases such as: model composition across tech- nical spaces, heterogeneous (meta)models alignment and keeping models in sync. In this paper, we introduce the models federation approach to conceptual interoperability that drives the development of our innovative modeling engine.