Vincenzo Stoico

Vincenzo Stoico serves as an Assistant Professor in the Department of Computer Science at the Faculty of Science, Vrije Universiteit Amsterdam (VU Amsterdam), with dual affiliation at the university's Network Institute. His academic profile centers on computational sustainability, specifically optimizing energy consumption in emerging AI-driven systems.

His research critically examines energy efficiency across machine learning infrastructure, with particular emphasis on large language models (LLMs). Key investigations include energy trade-offs in LLM-generated content delivery (on-device versus cloud), sustainable monitoring of concept drift in ML systems, and energy-aware self-adaptation mechanisms for robotics. This work bridges theoretical performance modeling with empirical validation in green computing contexts, addressing urgent environmental concerns in AI deployment.

Recent publications (2024-2025) demonstrate consistent focus on quantifying energy-performance relationships across diverse computing domains. His methodology combines controlled experimentation with architectural analysis, revealing significant energy savings through optimized LLM operations, concept drift detection frameworks, and behavior tree implementations in robotic systems. These contributions advance sustainable AI engineering practices while highlighting critical accuracy-energy tradeoffs.

Stoico actively collaborates with researchers including Ivano Malavolta and Patricia Lago across VU Amsterdam's Computer Science department and Network Institute. He contributes to academic instruction through the 'Green Lab' course, though specific advising relationships and grant funding details remain unreported in available sources.

His institutional engagement occurs primarily within VU Amsterdam's Network Institute ecosystem, which fosters interdisciplinary research on digital society challenges. Current projects likely integrate with the university's broader sustainability initiatives, though explicit lab affiliations or team structures are not documented in the provided materials.