Core Team
Riccardo Mansutti
Agentic AI in Energy
Master StudentAbout
Riccardo Mansutti is an electrical and electronics engineer whose work sits at the intersection of intelligent systems, critical infrastructure, and real-world deployment. With academic training at Politecnico di Milano and a specialization in sustainable smart grids for the energy transition, he brings a systems-level understanding of how complex physical networks operate under pressure. His background combines rigorous engineering foundations with an unusually practical orientation toward building technologies that can support high-stakes operational environments. Before joining the lab, Riccardo worked across both technical R&D and startup settings. Early research roles exposed him to sensing systems, control architectures, and embedded electronics, while later experience in an AI company focused on building agentic systems for enterprise workflows. Alongside this, he founded Astra, a mentorship initiative designed to help ambitious students develop through strong peer and mentor networks. Across these settings, a consistent theme runs through his work: designing intelligent systems that do not remain theoretical, but interact with real users, real constraints, and real infrastructure. At the lab, Riccardo focuses on how agentic AI can augment operators in power systems—one of the most consequential and operationally demanding domains of modern industry. His work reflects a broader ambition to make advanced AI dependable in environments where speed, interpretability, and safety matter at once.
Research Areas
Connect
Project
AI2GRID: Agentic Copilots for Power System Operations
AI2GRID is an agentic copilot for electric power system operators, designed to support decision-making in environments where complexity is rising faster than traditional tools can keep up. As renewable integration, distributed generation, and decarbonization targets reshape the grid, operators are increasingly required to manage unstable, high-dimensional systems under time pressure. AI2GRID acts as a ghost support layer that continuously monitors the real-time state of the grid, classifies operating conditions, detects anomalies or inefficiencies, and proactively proposes optimization workflows before an operator has to ask. Rather than functioning as a passive dashboard, it is built to surface context-aware, ready-to-evaluate recommendations while preserving strict human-in-the-loop control. From a scientific perspective, the project advances research in agentic AI for safety-critical cyber-physical systems. A central challenge is enabling an AI system to reason over dynamic, multivariate grid states and choose meaningful interventions under uncertainty, where incorrect actions could have severe physical and economic consequences. This raises deep questions in anomaly detection, state classification, action selection, and explainability in real time. AI2GRID also contributes to the study of human-machine teaming: how intelligent systems can become proactive enough to be operationally useful while remaining legible, controllable, and trustworthy to expert users working in high-stakes conditions. Utilities and grid operators worldwide are under pressure to modernize operations as they absorb more renewable volatility, electrification demand, and regulatory scrutiny. Yet much of their workflow still depends on fragmented software and manual expertise. AI2GRID points toward a new operational layer: an always-on decision support system that reduces response times, lowers the likelihood of human error, and helps teams manage increasingly complex grids with greater confidence. That makes it highly relevant for transmission operators, distribution operators, and utilities looking for deployable AI infrastructure that improves resilience, efficiency, and operational readiness without removing human oversight.
Other team members
Students
Interested in collaborating?
We are always looking for talented students, researchers and industry partners.
