Constantinos Tsakonas

Bio

Constantinos Tsakonas is a doctoral researcher at team HUCEBOT at Inria, Nancy, working on generative AI. He received his diploma from the Computer Engineering and Informatics Department (CEID) at the University of Patras, Greece in 2023. He worked as a researcher at the same university for almost three years, during and after his diploma, contributing to several AI projects. Notably, he worked as a robot learning researcher on the ‘Novel Optimization Methods for Autonomous Skill Learning in Robotics (NOSALRO)’ project, where he focused on learning representations, generative AI, autonomous skill discovery and learning, Quality-Diversity algorithms, and sim-to-real methods. Before that, he worked at the IoT Laboratory as a machine learning engineer and researcher, focusing on TinyML and AI-enabled embedded systems. He also contributed to Pfizer’s project, “Voice-based Diagnostics” (direct contract with IoT Laboratory), as a machine learning researcher where he worked on neural network architectures for audio data and state-of-the-art digital signal processing techniques.

He has co-authored multiple publications for international conferences and journals, and he serves as a reviewer on conferences considering Artificial Intelligence and Robotics, with the most notable being the NeurIPS, ICML, RA-L, and IROS. Moreover, he was a visiting researcher at TU Dresden, working on domain adaptation and cross-modal models for time series data.

Lastly, as an undergraduate, he won the local round of a European machine learning hackathon, where the problem formulation and solution review was carried out by Ernst & Young (EY), and later on, he represented the University of Patras at the final round in Milan, competing in a challenge provided by Infineon Technologies.