Salem Hamdani

– Developed and optimized CNN-based models with attention mechanisms for cardiovascular disease risk prediction, achieving 90% accuracy and enabling real-world clinical deployment. – Reduced model size by up to 60% and improved inference speed by 2x by applying advanced compression techniques using CompactifAI, with no significant loss in predictive performance. – Designed and deployed a scalable highway monitoring dashboard for Abertis on AWS, enabling real-time visibility across multiple highways.

Conducting research using satellite imagery and transfer learning to detect and analyze changes in ecological richness in agricultural fields, resulting in a publication at ICLR 2026, under the supervision of Evan Shelhamer.

– Contributed to Mindwaves V3 using Google Cloud Functions and Flutter, reducing AI API latency through prompt engineering while architecting multi-step orchestration pipelines for AI state management.. – Maintained Lintum and LegalTegrity, serving 500+ clients across both platforms ensuring high availability and continuity of service. – Built a PDF handling library from scratch in Vue.js and TypeScript, introducing advanced features including spot color support and 3D rendering.

– Migrated Doranco LMS from Symfony 3.4 to 5.4 for a platform serving 5,000+ users, modernizing legacy bundles and frontend stack (Twig, jQuery, Bootstrap), resulting in a 3× performance improvement. – Architected and delivered Okalee LMS, a multi-tenant SaaS platform built with React, TypeScript, and Symfony, successfully onboarding 20 clients post-launch.