PhD Students / PhD Alumni

PhD Students

• Liana Toderean Research: The research focuses on the orchestration of federated learning services across the smart grid computational continuum, where data, models, and computing resources are distributed across edge, fog, and cloud layers. The proposed framework coordinates federated learning workflows, participants, models, and computational resources while addressing data interoperability, privacy, governance, and efficient model aggregation. It further investigates adaptive orchestration mechanisms for distributed energy resources and virtual power plants, enabling federated AI models to support decentralized and privacy-preserving energy management.
• Gabriel Antonesi Research: The research focuses on advanced AI and Large Language Models for forecasting, flexibility estimation, and decision support in energy systems under complex and uncertain data conditions. It investigates hybrid architectures and data-driven approaches for forecasting and flexibility estimation with incomplete and noisy data. The work further explores agentic AI and LLM-based platforms for automating hypothesis testing, research workflows, and the development of AI-driven architectures and training pipelines for energy applications.
• Costel Frandes Research: Research focuses on computation orchestration for AI systems, addressing the dynamic allocation and coordination of heterogeneous computing resources across edge, fog, and cloud environments. It investigates resource-aware orchestration strategies for AI workloads, considering computational capacity, energy consumption, latency, communication overhead, and workload requirements. The work further explores adaptive and intelligent orchestration mechanisms that enable efficient deployment, scaling, and execution of AI models across distributed and heterogeneous computing infrastructures.
• Ioana Ramona Martin Research: developing trust quantification mechanisms for multi-agent knowledge, enabling AI agents to assess the reliability, provenance, and consistency of information shared across distributed environments. The research investigates intent-alignment verification across agents, ensuring that autonomous agents can identify compatible objectives, constraints, and decision policies before cooperation. It further develops coordination protocols and empirical validation frameworks for trustworthy multi-agent collaboration in complex physical systems.
• Ofrim Vasile Research: Investigates accountable execution in compound AI systems, focusing on the gap between optimizing proxy objectives and verifying whether an execution fulfills its intended goal. Will address aspects related to the formalization of user intents as verifiable constraints over quality, cost, latency, governance, and execution traces, while refining end-to-end requirements into module-level obligations and resolving conflicts among them. The expected outcome is a foundation for future AI orchestration systems capable of providing runtime evidence that executions satisfy their specified objectives and constraints.

Master's Thesis Supervision

• Mihai Daian Research: Research focuses on federated learning and Large Language Models for decentralized AI across resource-constrained edge environments, where limited computational, memory, communication, and energy resources challenge conventional training and inference approaches. It investigates efficient federated orchestration, model compression, adaptive participation, and privacy-preserving techniques to enable the collaborative training and deployment of LLM-based models across heterogeneous edge devices.

PhD Alumni

• Dan Mitrea Thesis: Blockchain-enabled decentralization mechanisms for Smart Grids facilitating privacy, energy cooperation, and edge orchestration (2025). Defense announcement · Summary (PDF)
• Gabriel Arcas Thesis: Contributions to the Development of Orchestration Solutions for Computational Continuum over the Smart Grid (2025). Defense announcement · Summary (PDF)