CV
Azzeddine Tiba
Summary
Scientific Machine Learning Research Engineer at Michelin. PhD from CNAM Paris on machine learning strategies for accelerating numerical simulations of fluid-structure interaction.
Education
- Applied Mathematics / Scientific Machine Learning2024-11-01CNAM - Conservatoire National des Arts et Métiers
- Scientific Machine Learning2023-07-31CEMRACS 2023, CIRM
- Mechanical Engineering2021-08-31ENSAM - École Nationale Supérieure d'Arts et MétiersGPA: Silver Medal - Rank: 83/1180
Work Experience
- Computational R&D Engineer2024-12-01 -MichelinPart of a team developing a large-scale industrial finite element framework designed for massively parallel HPC environments.
- PhD Student - Research Engineer2021-11-01 - 2024-10-31M2N, CNAM - Michelin, AltairMachine Learning-based Reduced Order Models (ROMs) for fluid-structure interaction (FSI). Published 2 journal articles and multiple international conference communications.
- Visiting PhD Student2024-02-01 - 2024-03-31Esteco (Previously Optimad)Explored stability properties of data-driven reduced order models.
- Substitute Teacher2022-09-01 - 2024-01-31CNAMTaught practical work on Numerical Methods, Fluid Mechanics and Functional Analysis.
- Simulation Research Engineer (Master's Thesis)2021-03-01 - 2021-08-31Dassault SystèmesData-Driven Computational Mechanics (DDCM). Studied numerical convergence and extension to non-elastic and multiscale problems. Extended the approach using manifold learning and noisy optimization techniques.
- Simulation Software QA Intern2020-06-01 - 2020-08-31Coventor Inc. (Lam Research)Wrote tests for meshing features of MEMS+, involving numerical analysis and electromechanical modeling.
Skills
Programming
- Python
- C++
- C
- Matlab
Scientific Computing
- PETSc
- KratosMultiphysics
- FEniCS
- scikit-fem
- NumPy
- PyTorch
- scikit-learn
- PyVista
- Paraview
Tools
- Git
- Linux
- LaTeX
Scientific Expertise
- Linear Algebra
- Dynamical Systems
- Nonlinear Solvers
- Computational Mechanics
- Finite Element Method
- Multiphysics Coupling
- Reduced-Order Modeling
- Manifold Learning
Publications
- Machine-learning enhanced predictors for accelerated convergence of partitioned fluid-structure interaction simulations2025Computer Physics CommunicationsDevelops ML-enhanced predictors to accelerate convergence of partitioned FSI coupling algorithms.
- Non-intrusive reduced order models for partitioned fluid–structure interactions2024Journal of Fluids and StructuresPresents non-intrusive reduced order models for partitioned fluid-structure interaction simulations.
- Machine Learning to Accelerate Fluid-Structure Interaction Simulations2025FSSIC 2025 Symposium ProceedingsAccepted conference paper on ML methods to accelerate FSI simulations.
- Dynamical Data-Driven Model Order Reduction for nonlinear Fluid-Structure Interaction problems202225è Congrès Français de MécaniqueConference paper on dynamical data-driven model order reduction for nonlinear FSI.
Presentations
- SPARCL Workshop2026SPARCL WorkshopParis, FranceTalk
- 37ème Séminaire de mécanique des fluides numérique202537ème Séminaire de mécanique des fluides numérique (SMAI-GAMNI)IHP, ParisTalk
- ICCE Conference2024ICCE ConferenceDarmstadt, GermanyInvited talk
- CEACM S4ML Conference2024CEACM S4ML ConferencePrague, Czech RepublicInvited talk
- ECCOMAS2024ECCOMASLisbon, PortugalInvited talk
- ERCOFTAC2023ERCOFTACToulouse, FranceInvited talk
- SIA Simulation Numérique2023SIA Simulation NumériqueChamps-sur-Marne, FranceTalk
- 1st Aria Workshop20231st Aria WorkshopBordeaux, FranceInvited talk
- CANUM2022CANUMEvian-les-Bains, FranceInvited talk
Teaching
- Numerical Methods, Fluid Mechanics and Functional Analysis2022CNAM, ParisRole: Substitute Teacher (Practical Work)Taught practical work on Numerical Methods, Fluid Mechanics and Functional Analysis.
Portfolio
- ROM_AMPortfolioA Non-intrusive Reduced Order Modeling package using data-driven methods and Machine Learning.
- FeCLAPPortfolioA finite element package for the simulation of laminated composite plates in elasticity and plasticity.
- ROM-FOM Coupling WorkshopPortfolioTutorial-like course on non-intrusive coupling between Reduced Order Models and Classical models.
Languages
- EnglishTOEIC: 985/990
- FrenchNative
- ArabicNative