Projet de fin d'étude : Intelligent Deployment Optimization of AI Solutions
Etudiant : EL HAMDOUNI YOUSSEF
Filière : Master Web Intelligence et Sciences des Données (WISD)
Encadrant : Pr. SABRI ABDELOUAHED
Annèe : 2025
Résumé : The project involves comparing a language model (LLaMA 3.2 3B) with a traditional model (XGBoost) for a resume classification task. The evaluation focuses on performance, cost, and deployment feasibility in a cloud production environment (AWS SageMaker), using an Infrastructure-as-Code approach via Terraform. The results show that preprocessing strategies must be adapted to the type of model, and that traditional models remain advantageous in low-resource contexts.