Digital Twins (DT) constitute complex software systems that need to be continuously modified/updated to meet evolving user requirements and priorities, and support continual improvement. Because they aim at monitoring and improving different system aspects, their development requires the collaboration of people from different domains of expertise, e.g. the development of a DT for build assets may involve the collaboration of experts in software engineering, thermal comfort, air quality, and energy consumption. Consequently, DTs need to be engineered to enable the fast and secure integration and deployment of new code, the systematic and iterative evolution of their components, and the independent development of different system aspects by people with different expertise.

In this paper, we present a DevOps framework for the engineering of DTs for build assets. The framework is being developed in the context of the ETS GRIDD lab, a multi-function room at ÉTS, used to develop, test, and validate it. We focus on two main aspects: the microservices architecture of the DT software; and the DevOps infrastructure used to support the development, continuous integration, and continuous delivery of the DT software. We also discuss lessons-learned, challenges, and next steps related to the development and evolution of the framework.