At IIAS, CHU DE REIMS, I work on Predibilia, using NLP to extract medical histories from clinical reports and normalize them with CIM-10 coding. I developed approaches using Python, regular expressions, and medical terminology resources, as well as an AI agent built with two LLMs.
For Predibilia, I coordinated two offline agent instances to reduce processing time and set up continuous saving of their outputs. I co-authored an English-language article submitted for publication and created Argilla interfaces for doctors to review NLP results.
I also worked on StaphScore, creating a web interface for biologists to annotate microscopy images for a planned binary image classifier. I developed a Python tool to collect image tiles and polygon crop annotations from Cytomine, and used Cellpose for segmentation.
Before this, I completed a doctorate in pure mathematics at UNIVERSITÉ DE REIMS, researching quantum deformations and projective representations. I co-authored an article published in Letters in Mathematical Physics and taught pure mathematics to first-year computer science students.

