At CARTE ASSURANCES, I developed and deployed an interactive platform for real-time insurance premium simulation. I also created automated Python file-reconciliation algorithms to detect inconsistencies across Excel files and reduce manual checking time.
At CARTE ASSURANCES, I analyzed and cleaned data on contract structures, claims ratios, and actuarial pricing models.
For my master’s thesis at Université Paris Dauphine - PSL, I studied nonparametric regression estimators using random projections and ANOVA truncation. I evaluated the approach against Random Forest, Gradient Boosting, and MLP models using cross-validation, test functions, and real datasets.
At the GEN AI For GOOD hackathon, I developed a multimodal chatbot combining vision, text, and a RAG architecture for plant disease diagnosis and biological treatment recommendations. At Hack for Care and Culture with AI, I optimized machine-learning models for early Parkinson’s diagnosis.

