The changes facing the electricity sector are part of a process that is more gradual than disruptive, in which technologies such as virtualization and the application of algorithms, whether deterministic or AI-based, are two of the catalysts that will drive the transition from a centralized energy system to a more decentralized, local, and efficient one. AI technologies and their algorithms are at an essential level of development when trained and executed at scale in the Cloud. However, there are other technological trends for distributing algorithms that leverage the distributed computing capacity of local equipment at the operational level: Federated Learning, AI at the Edge, and Virtualization. But they also have high computing and memory requirements. There is significant scope for research to explore the limits of AI on devices with limited computing resources (Resource-Constrained Devices, RCD). The technological response is an evolution of microcontroller-based devices with integrated technologies, optimizing their performance, security, and processing capacity in critical applications. The combination of these technologies with modern microcontrollers not only improves computational efficiency and security but also enables the development of intelligent, reliable, and high-performance embedded systems in demanding sectors such as SG. In this context, the overall objective of the FIRMAR project is to research architectures that enable the secure embedding of advanced algorithms for electrical networks in Resource-Constrained Devices (RCDs). The architectures and performance of RCD platforms will be analyzed in relation to the functional requirements of electrical networks and the feasibility of transferring cloud deployment models to Edge-RCD. RCD development platforms, including System-on-Chip (SoC), will be investigated for their ability to meet cybersecurity requirements to ensure resilience, support AI techniques, and evaluate green algorithms. Finally, use cases for application in the electricity sector, in particular to electrical networks, that validate the research carried out in the project are already being identified and will be explored in the project.