About Hugo
[EN] About me
🔎 I help companies and research teams turn complex scientific models into fast, actionable tools.
👨🏼💻 As a freelance, I focus on two key bottlenecks:
- Speeding up long numerical simulations from hours to seconds using surrogate models
- Calibrating complex physical models with thousands of parameters automatically from data
I combine Machine Learning and Physics to deliver solutions that are both fast, reliable and usable in practice.
🎓 During my PhD, I developed a Physics-Informed Machine Learning algorithm capable of calibrating a fluid dynamics model with 1000+ parameters in 30 seconds, instead of 5 hours with standard CFD methods, enabling real-time use and large-scale analysis.
[FR] À propos de moi
🔎 J’aide les entreprises et équipes de recherche à transformer des modèles scientifiques complexes en outils rapides et exploitables.
👨🏼💻 En freelance, je me concentre sur deux verrous majeurs :
- Accélérer les simulations numériques longues de plusieurs heures à quelques secondes grâce à des modèles de substitution (surrogate models)
- Calibrer des modèles physiques complexes avec des milliers de paramètres automatiquement à partir des données
Je combine le Machine Learning et la physique pour proposer des solutions à la fois rapides, fiables et utilisables en pratique.
🎓 Durant ma thèse, j’ai développé un algorithme de Physics-Informed Machine Learning capable de calibrer un modèle de mécanique des fluides avec plus de 1000 paramètres en 30 secondes, contre 5 heures avec des méthodes CFD classiques, rendant ainsi possible une utilisation de ce modèle en temps réel et des analyses à grande échelle.
French
Native or bilingual
English
Fluent
Remote only
Primarily works remotely
Experience
- FreelanceFreelance in Scientific Machine Learning (SciML) & Data ScienceTECHJanuary 2026 - Today (5 months)Montpellier, FranceI help companies and research teams turn complex scientific models into fast, actionable tools.To achieve this, I focus on two key bottlenecks:
- Speeding up long numerical simulations from hours to seconds using surrogate models
- Calibrating complex physical models with thousands of parameters automatically from data
I combine Machine Learning and physics to deliver solutions that are both fast, reliable and usable in practice.To illustrate this, during my PhD, I developed a Physics-Informed Machine Learning algorithm capable of calibrating a fluid dynamics model with 1000+ parameters in 30 seconds, instead of 5 hours with standard CFD methods, enabling real-time use and large-scale analysis. - Institut de Mathématiques de ToulousePhD student in Applied Mathematics on Physics-Informed Machine Learning for flood simulationRESEARCHJanuary 2022 - Today (4 years and 5 months)Toulouse, FrancePhD in Applied Mathematics | Physics-Informed Machine Learning for Flood ModelingI’m developing Physics-Informed Neural Networks (PINNs) and other Physics-Informed Machine Learning architectures for high-dimensional data assimilation applied to the Shallow Water Equations (SWE) and other hydraulic models, using PyTorch.My research bridges Machine Learning and physics-based modeling to improve flood simulations, parameter estimation, and model calibration in complex spatio-temporal domains.This work lies at the intersection of Scientific Machine Learning (SciML), fluid dynamics, and applied mathematics, focusing on inverse problems, numerical simulations, and the design of efficient neural architectures for the solution of partial differential equations (PDEs).Supervised by J. Monnier (INSA Toulouse / IMT), P.-A. Garambois (INRAE Aix-en-Provence), and R. Bouclier (INSA Toulouse / ICA).📄 Related publication:"Spatially-distributed parameter identification by physics-informed neural networks illustrated on the 2D shallow-water equations", published in IOP Inverse Problems.
- CNESConsulting Engineer in Cryogenic Fluid Systems for THEMISAVIATION AND AEROSPACEJanuary 2020 - January 2021 (1 year)Paris, FranceConsulting Engineer for a mission consisting in the modelization on EcosimPro/ESPSS of Cryogenics Systems for ground operations on the reusable launcher demonstrator Themis, developed by ArianeWorks.Missions achieved :• Functional modelization with EcosimPro/ESPSS of the Themis ground and board fluid systems.• Simulation and sequence optimization for the chill-down, filling, pressurization, draining and depressurization procedures of the Themis ground and board fluid systems.• Development of the Themis ground fluid systems architecture with a Piping & Instrumentation Diagram (PID) and an Interface Control Document (ICD) between ground and board fluid systems.
Recommendations
Be the first to recommend Hugo
Help this freelancer shine by sharing your experience working together.
These freelancer profiles also match your criteria
Agatha Frydrych
Backend Java Software Engineer
4.7
(3)
2
Baptiste Duhen
Fullstack developer
4.6
(4)
5
Amed Hamou
Senior Lead Developer
4
(2)
7
Audrey Champion
Web developer
4.3
(3)
4
Education
- PhD, Applied Mathematics on Physics-Informed Machine Learning for flood simulationInstitut national des Sciences appliquées de ToulousePhD student in Applied Mathematics on Physics-Informed Machine Learning for flood simulation. Currently investigating Physics-Informed Neural Networks (PINNs) for high-dimensional Data Assimilation applied to the Shallow-Water Equations (SWE), using Pytorch.
- Specialization Degree (MSc level), Aeronautics and Aerospace EngineeringCentraleSupélec2020EPFxCentraleSupélec Program of Academic Excellence, Specialization Degree, attending the "Ingénieur Centralien" Curriculum final year in the "Mechanics, Aeronautics, Aerospace" Specialization. Advanced courses in Fluid Dynamics.