Paul Scharnhorst

Paul Scharnhorst is a researcher at École Polytechnique Fédérale de Lausanne (EPFL), focusing on energy systems, data-driven control, and building automation. His work bridges theoretical advancements in machine learning and practical applications in renewable energy integration and demand response. He has contributed to the development of uncertainty-aware modeling frameworks and open-source tools for benchmarking building controllers.

Education: Completed a doctoral thesis in 2024 titled Quantifying the Unknown: Data-Driven Approaches and Applications in Energy Systems under supervision of Colin Neil Jones and Baptiste Schubnel at EPFL.

Research emphasizes robust control methodologies, kernel-based learning, and energy flexibility coordination in decentralized systems. His tools, such as the Energym library, facilitate standardized testing of building control strategies using simulation models from EnergyPlus and Modelica.

Key contributions include uncertainty quantification in battery models for buildings, deterministic error bounds in kernel regression, and learning-based predictive control with safety guarantees. Collaborations involve interdisciplinary teams addressing challenges in smart grid integration and building energy management.