The FINALITY project will be submitted to the call Marie Skłodowska-Curie Action (MSCA) on Doctoral Networks (DN) in late 2023. The MSCA proposal will define an academic DN with industrial participation. The aim of FINALITY is to form a novel AI curriculum for engineering researchers able to fill the current lack of competences in safe learning. In fact, wafe learning is an emerging development in learning theory and it is currently under-represented in the background of engineering researchers. Indeed, AI is expected to empower intelligent resource management for socio-technical systems, i.e., where technological systems require human decisions for the allocation of resources which can be supported by AI. FINALITY will explore methods such as Reinforcement Learning, Online Convex Optimization and Federated Learning: while they have been proven efficient in various fields in recent years, they still have significant limitations before having an impact. Firstly, in fact, resources allocation in socio-technical systems is subject to multiple constraints as those dictated by the system structure, e.g., a network structure, or arising in many cases from human factors such as regulatory constraints and/or the current practices which are adopted in the industrial domain. This is a particularly challenging setting for the operations of AI algorithms, which are not designed natively to account for limitations in the state and action space, and need to be rendered aware of such constraints. Nowadays, the literature and the practice to this regard is not yet mature and requires specific effort. The researchers engaged in the FINALITY DN will develop new methodological tools focusing on the integration of research areas in which the principal investigators of the consortium are currently active and knowledgeable. Those include constrained and delayed MDP theory and their application to safe reinforcement learning and recent advancements on online convex optimization and federated learning. The consortium has a consolidated record of collaborations and joint publications and it is formed by top notch experts in both the theoretical and the applied domains. The DN will extend the competence basis at the EU level with scientific effort well aligned to the national and EU priorities in the field of AI research. The MRSEI action will be pursued with a series of plenary and specific meetings, during an initial consolidation phase where the consortium will be completed and the research lines will be further detailed according to a tight agenda towards the submission of the proposal.