MALEK ZITOUNI

Generative AI Engineer Intern

• Job Details:Architected and deployed vision-language model (VLM) system for automated KYC document processing using GPT-4V and custom fine-tuned models, reducing manual review time by 40% and achieving 98% compliance accuracy. Engineered dynamic schema mapping pipeline with transformer-based document understanding, enabling real-time processing of 200+ daily credit applications across heterogeneous financial document formats. Implemented MLOps pipeline with automated model retraining, A/B testing framework, and performance monitoring, improving model drift detection by 60%.

3D Computer Vision AI Intern

• Job Details:Conducted comprehensive evaluation of state-of-the-art 3D deep learning architectures (PointNet++, RandLA-Net, Point Transformer) for automated BIM generation from point clouds, achieving 92% precision in structural element classification. Designed and implemented end-to-end ML pipeline with data preprocessing, feature engineering, and post-processing optimization, reducing manual correction time by 25% across 15+ concurrent construction projects. Developed custom loss functions and data augmentation strategies for sparse 3D point cloud data, improving model generalization by 18%.