Summary
Ahmed Mazari is a Staff AI Engineer and applied scientist with nine years of experience specializing in geometric deep learning for fluid dynamics, neural operators, and PDE-based simulation at Ansys/SimAI. He holds a PhD from Sorbonne Université focused on graph convolutional networks and multiple kernel learning for video action recognition, and transitioned his expertise from computer vision to large-scale mesh and equivariant models for CFD. His work bridges research and production: developing mesh/graph neural networks, multi-scale representations, uncertainty quantification, and hybrid simulation approaches that accelerate numerical solvers. Ahmed has a track record of high-impact internships and R&D roles, including OCR and deep generative model projects, demonstrating both theoretical depth and practical engineering. Based in Paris, he combines rigorous academic training with industry-scale deployments, often tackling the interface of physics, geometry, and scalable ML. An under-the-radar strength is his early work on optimal consensus protocols and distributed systems, which informs robust, fault-tolerant design in complex simulation pipelines.
9 years of coding experience
Master’s Degree, Artificial Intelligence, Master’s Degree, Artificial Intelligence at Paris-Sud University (Paris XI)
Master’s Degree, Artificial Intelligence and machine learning for data science, Master’s Degree, Artificial Intelligence and machine learning for data science at Université Paris Cité
Master’s Degree, Distributed computing and networking, Master’s Degree, Distributed computing and networking at University of Bejaia - Algeria
Doctor of Philosophy - PhD, Graph Convolutional Neural Networks and Multiple Kernel Learning for Action Recognition in Videos, Doctor of Philosophy - PhD, Graph Convolutional Neural Networks and Multiple Kernel Learning for Action Recognition in Videos at Pierre and Marie Curie University
Kabyle, English, French, berber, formal and algerian arabic