Ayoub EL HOUDRI

Providing expert services in data science, machine learning, and data collection for clients across various industries, including insurance, marketing, retail, and software development. Profile: malt.fr/profile/ayoubelhoudri

- Implementation of a customized generative model for 3D lung segmentation, utilizing architectures like U-Net. - Finetuning the SAM (Segment Anything Model) on CT data for precise lung segmentation. - Processing diverse medical images, including MRI and CT scans. - Designing a pipeline for tumor detection within lung volumes. - In-depth analysis of tumor malignancy based on their structure and texture. - Monitoring computing clusters used for model training.

- Implementation of an effective document preprocessing system, incorporating metadata verification and hashing techniques to ensure compliance. - Development of a tabular information detection model, followed by finetuning the LayoutParser OCR model for character recognition within the table. - Solution deployment as an internal API for accessibility within the organization. - Integration into an MLOps approach to manage the model's lifecycle, with continuous monitoring post-tests.

- Compilation of a bibliography on shape reconstruction models from images using Graph Neural Networks (GNN). - Augmentation and processing of synthetic 3D data, including PointCloud, Mesh, and Voxel types. - Finetuning of the Pixel2Mesh model for shape reconstruction from a single image. - Deployment of the model in production, integration into a platform dedicated to clients.