Martin Roa-Villescas

Martin Roa-Villescas is a Doctoral Candidate and Teaching Assistant in the Department of Electronic Systems within the Electrical Engineering school at Eindhoven University of Technology (TU/e). He also holds an external position as a Docent (Lecturer) at Fontys University of Applied Sciences since August 2023.

His educational background includes a B.Sc. in Electronic Engineering from the National University of Colombia (2010) and an M.Sc. in Embedded Systems from TU/e (2013). Between 2013 and 2018, he worked as an embedded software designer at Philips Research in Eindhoven.

Roa-Villescas' research focuses on probabilistic graphical models, Bayesian machine learning, and tensor network applications. His work spans both theoretical advancements in probabilistic inference algorithms and practical implementations in open-source software. He has made significant contributions to optimizing message passing algorithms and developing efficient execution schedules for Bayesian inference.

His recent publications demonstrate a clear trajectory toward improving the scalability and efficiency of probabilistic inference methods, with applications spanning audio signal processing, robotics, and hearing aid technology. The development of TensorInference, a Julia package for tensor-based probabilistic inference, represents his commitment to open-source tools that advance the field.

As a researcher involved in the STW Zero Autonomous Acoustic Systems project (2017-2024), he has contributed to advancing acoustic sensing technologies through probabilistic modeling approaches. His work aligns with UN Sustainable Development Goals related to technology innovation and accessibility.