At Université du Québec à Chicoutimi, I contributed to an automatic cost estimation system for carpentry from construction plans. I prepared training datasets, annotated floor-plan images, and trained and evaluated YOLO and RT-DETR models to detect construction elements.
I also developed scripts to convert annotations into YOLO format and integrated an AI backend into Label Studio for automatic pre-annotation. I tested SAHI with RT-DETR and validated predictions to improve annotation quality and consistency.
In a 2023–2024 postdoctoral project with Rio Tinto, I analyzed greenhouse-emissions monitoring data from aluminium production using statistical calculations and signal segmentation. My work produced results using descriptive statistics.
As a lecturer at Université du Québec à Chicoutimi, I teach mathematics courses including advanced calculus, real analysis, and vector and matrix algebra. Earlier, I delivered training in data analysis, artificial intelligence, and statistics at SKILLART Consulting.
