Valentin Caries
Research Engineer @ IFPEN · PhD · Scientific Python | CFD | Simulation
IFP Energies nouvelles (IFPEN)
Biography
I am Valentin Caries, a research engineer at IFP Energies nouvelles (IFPEN), where I work on multiphase pump systems. I hold a PhD in aerodynamics from École Centrale de Lyon (2025), carried out in partnership with Safran Aircraft Engines. My work sits at the meeting point of fluid mechanics, numerical methods and software engineering: I design physical models, turn them into fast and maintainable code, and validate them against high-fidelity simulations and experiments.
During my PhD I developed a low-order, three-dimensional aerodynamic solver that predicts the flow around compressor and fan rotors — including the notoriously difficult tip-leakage flow — in seconds instead of hours, making it usable for early-stage design exploration. I have worked across the full computational spectrum, from wall-resolved Large Eddy Simulation (LES) on HPC clusters to reduced-order panel and vortex-lattice methods, and built the data and post-processing tooling that ties it all together.
Alongside my research, I am open to selected freelance and consulting work in scientific Python, CFD and simulation, data engineering, and machine learning for engineering. If you need someone who understands both the physics and the code, let’s talk.
- Scientific Python development
- Computational Fluid Dynamics (CFD)
- Multiphase flow & turbomachinery
- Reduced-order & multi-fidelity modeling
- Machine learning for engineering
PhD in Aerodynamic Modeling for Turbomachinery, 2022 – 2025
École Centrale de Lyon
Advanced MSc in Aerospace Propulsion, 2020 – 2021
ISAE-SUPAERO
MSc in Mechanical Engineering (dual degree), 2019 – 2020
École Centrale de Lyon
Engineering degree, Mechanical Engineering, 2017 – 2020
Polytech Lyon
Skills & Services
I help engineering and research teams turn complex physics into fast, reliable, well-tested software and turn raw simulation data into insight. Open to selected freelance and consulting missions.
🐍 Scientific Python
Numerical methods, solvers and reusable packages. Clean, documented, tested code with NumPy/SciPy, performance profiling and packaging.
🌀 CFD & Simulation
RANS & LES workflows, reduced-order and multi-fidelity aerodynamic models, mesh handling, solver setup, verification & validation.
⚙️ Data Engineering & HPC
Post-processing pipelines, large dataset handling, job automation on HPC clusters (SLURM), reproducible and parallel workflows.
Experience
Research and scientific development on multiphase pump systems.
- Numerical modeling and CFD of multiphase flows.
- Experimental campaigns and data analysis.
- Scientific Python development for simulation, data processing and analysis.
Multi-fidelity modeling of the tip-leakage flow for an axial compressor rotor in compressible flow. PhD defended on 11 June 2025.
- Designed and implemented a 3D low-order aerodynamic solver (panel + vortex-lattice methods) in Python, reducing rotor flow predictions from hours to seconds.
- Validated the models against RANS simulations and experiments; published in peer-reviewed journals and international conferences.
- Built the supporting data processing and visualization tooling.
- Developed and automated RANS simulation methodologies.
- Delivered Python tooling to streamline pre- and post-processing.
Wall-Resolved Large Eddy Simulation (WRLES) of 2.5D propeller configurations.
- Ran high-fidelity LES on HPC clusters and developed the post-processing methodology in Python.
- Built a Python model for Shell Eco-marathon race-strategy optimization (vehicle dynamics, energy management).
Projects
Selected engineering & software projects
Recent publications
A low-order 3D Hybrid Panel Method predicts the flow around low-pressure compressor rotors and feeds dedicated tip-leakage models — in under a minute per case. Benchmarked against RANS on the LMFA NACA65 rotor, it reproduces rotor aerodynamics well and tip-leakage trends accurately for small tip gaps, supporting fast early-stage design.
🏆 IJTPP Editors’ Choice. This paper introduces a low-order, three-dimensional method for predicting inviscid flow around shrouded fans, aimed at early-stage design. By combining the vortex lattice and panel methods with a mixed boundary condition, it allows efficient exploration of design options. The method also models tip-leakage flow through an iterative algorithm. A periodicity condition is validated, reducing computational demand, with calculations completed in under a minute. The method agrees well with RANS for mean flow and tip-leakage characteristics, though some discrepancies arise at lower mass flow rates.
Let’s work together
Freelance & consulting — scientific Python, CFD, data, ML
Have a project in mind, or a problem at the boundary of physics and code? Get in touch.