Introduction
About Me
I am a PhD researcher at Université Paris-Saclay (IBISC Lab & CEA), working under the supervision of Dr. Martyna Poreba, Dr. Michal Szczepanski, and Prof. Samia Bouchafa. My research sits at the intersection of computer vision, large multimodal models, and resource-constrained deployment.
My core focus is making Vision-Language Models (VLMs) and Vision-Language-Action (VLA) models efficient enough to run on edge hardware — without retraining. I design training-free methods that analyse and reduce redundant visual tokens before they reach the LLM backbone, with a particular emphasis on deployment on the NVIDIA Jetson Orin for autonomous systems.
I am a Demythif.AI Fellow under Institut DATAIA, co-funded by the Marie Skłodowska-Curie Actions (MSCA) and the European Union — a programme dedicated to responsible, human-centred AI.
