Summary
Umberto Mele is an AI & Data Scientist with eight years of experience applying advanced machine learning to energy systems, currently developing production and consumption forecasting models at BKW AG in Switzerland. A PhD candidate at IDSIA/USI, he bridged deep learning and reinforcement learning with combinatorial optimization—designing heuristic solvers for TSP—which gives him a rare mix of theoretical rigor and practical algorithm design. His work spans diffusion models for consumption profiling, graph- and ontology-based RAG systems, and exploratory AI approaches for price and weather prediction to support the energy transition. He combines strong academic credentials (PhD research, top-ranked Master’s in Data Science) with hands-on industry delivery in regulated utility settings. Based in Lugano, he leverages domain knowledge of climate-weather-energy interactions to translate novel AI methods into operational forecasts. An interesting thread through his career is using cutting-edge generative and RL techniques not just for novelty but to make energy planning more robust and interpretable.
8 years of coding experience
Summer school, Computer Science, Summer school, Computer Science at Eidgenössische Technische Hochschule Zürich
Corso di Master Universitario di 2° livello-CMU2, Finance and Financial Management Services, Corso di Master Universitario di 2° livello-CMU2, Finance and Financial Management Services at MeliusForm Business School
Doctor of Philosophy - PhD, Artificial Intelligence, Doctor of Philosophy - PhD, Artificial Intelligence at USI Università della Svizzera italiana
Master's degree, Data Science, 110/110 con lode, Master's degree, Data Science, 110/110 con lode at Sapienza Università di Roma
Italian, Portuguese, English