Current position
Mohsen Seyedkazemi Ardebili is a Research Fellow (postdoctoral researcher) at the Department of Electrical, Electronic and Information Engineering "Guglielmo Marconi" (DEI), University of Bologna, in the research group of Prof. Andrea Bartolini and Prof. Luca Benini. His work brings artificial intelligence into the monitoring, prediction and autonomous management of high-performance computing (HPC) and cloud-native infrastructure.
Research
His research develops AI systems that operate large-scale computing infrastructure rather than only describe it, along three connected lines:
- Autonomous agents for HPC and Kubernetes operations. LLM-orchestrated multi-agent architectures that execute cluster operations behind an explicit human-in-the-loop approval gate, with runtime tool synthesis and role-based access control, together with benchmarks for evaluating such agents in operational settings (KubeIntellect, AOBench).
- Anomaly detection and predictive monitoring for Tier-0 datacentres. Deep-learning, statistical and graph-based frameworks for thermal and system anomaly prediction on exascale-class supercomputers (HazardNet, ThermADNet, GRAAFE), validated on CINECA's Marconi100, together with the M100 ExaData and PM100 open datasets.
- Digital twins, MLOps and sustainable computing. Real-time system modelling across the device-edge-cloud-HPC continuum, end-to-end MLOps platforms for AI in production, and power- and carbon-aware workload placement.
European research projects
He is proposal writer and UNIBO technical lead for MLOps and AI production in the EuroHPC JU project SEANERGYS (Software for Efficient and Energy-Aware Supercomputers, GA 101177590). He was UNIBO technical lead in the Horizon Europe project DECICE (GA 101092582), where he led the dynamic digital-twin task and the UNIBO cloud-robotics use case, and has contributed to Graph-Massivizer (GA 101093202), The European PILOT (GA 101034126), REGALE (GA 956560) and the European Processor Initiative (EPI SGA1, GA 826647).
Teaching and supervision
Since 2021 he has been project supervisor for the course 87195 - Lab of Big Data Architectures M, second-cycle degree programme in Electronic Engineering, and has contributed to 91279 - Project Work in Architectures and Platforms for Artificial Intelligence, second-cycle degree programme in Artificial Intelligence. Course topics have progressed from large-scale data analysis and deep learning for time-series anomaly detection to agentic AI application design, multi-agent orchestration and the Model Context Protocol. He co-supervises two Ph.D. students and has co-supervised five M.Sc. theses at DEI.
Academic service
Programme committee member for PDP 2025 and PDP 2026 (Euromicro International Conference on Parallel, Distributed and Network-Based Processing) and for AsHES 2026; co-organiser of the REGALE EU project meeting at the University of Bologna (2023). He reviews for IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, Engineering Applications of Artificial Intelligence, Sustainable Computing: Informatics and Systems, Journal of Grid Computing and Scientific Reports (Nature Portfolio), and for the DATE, SC, ACM Computing Frontiers, PDP and AsHES conferences.
Education
Ph.D. in Electronics, Telecommunications and Information Technology, University of Bologna (2018-2022). Supervisor: Prof. Luca Benini; co-supervisor: Prof. Andrea Bartolini. Thesis topic: design, analysis and management of high-performance computing systems.
M.Sc. in Telecommunications Engineering, University of Bologna (2015-2018). Thesis: Software-Defined Networking applied to survivable C-RAN, an intent-based implementation. Supervisor: Prof. Carla Raffaelli.
B.Sc. in Electrical Engineering, Islamic Azad University, Ardabil, Iran (2002-2006).
Professional experience before academia
Seven years of enterprise IT and network infrastructure experience, most recently as IT & Network Administrator of the Sabalan combined-cycle power plant (1,000-1,440 MW) in Ardabil, Iran, with full responsibility for the facility's enterprise IT and networking infrastructure across eleven operational zones, including a fibre-optic campus backbone, Windows Server and Active Directory environments, VMware virtualisation, security, backup and disaster recovery.
Languages
English (professional working proficiency), Italian (A2), Persian, Azerbaijani and Turkish (native).
Online profiles
ORCID: 0000-0002-1166-6559 | Google Scholar | GitHub | LinkedIn | Personal website: https://mskazemi.com/
