FLYING WHALES is an extraordinary industrial adventure.
We are developing an air transport solution to support the economic development of landlocked regions and significantly reduce the environmental footprint of cargo transport.
FLYING WHALES has formed a consortium to design, build on three continents and operate worldwide a fleet of LCA60T airships capable of loading and unloading up to 60 tons of cargo (wood, wind turbine blades, containers, mobile hospital, etc.) in hover flight, with a particularly low environmental footprint. This transport system is unique in the world. FLYING WHALES is also developing FLYING WHALES SERVICES, the operating company for this fleet, which will combine the skills of an airline and an airport company.
At FLYING WHALES, we have a highly ambitious vision that today brings together over 200 pioneers, a powerful international shareholder base and some 50 partner companies. Pioneers of the 21st century, men and women who contribute their desire, passion and know-how every day to give birth to a new transport solution, but also to an idea. The idea that companies must be able to continue creating links in order to grow.
This is FLYING WHALES' mission: to open up isolated regions of the world, while helping to reduce CO2 emissions and transport infrastructure.
You want to help us write a page in the history of aeronautics and the economic development of isolated regions? Here's the pioneer we're looking for:
WHO ARE WE LOOKING FOR?
Within the Loads & Aeroelasticity team, you will contribute to building and validating a proof-of-concept tool that demonstrates the added mass effects of the surrounding air on the LCA60T airship during a Safe Forced Landing (SFL) scenario. The tool will be developed with significant LLM assistance (Claude API), providing direct insights for FLYING WHALES’ future R&D strategy around AI-assisted scientific computing.
YOUR MISSIONS:
You will contribute to:
Phase 1 – Onboarding: tools, theory & application
Read and understand the Le Mestre (2022) PhD thesis, focusing on the BEM formulation, FEM structural model, coupling architecture, and added-mass derivation (Claude used as a reading & explanation assistant).
Get hands-on with the existing BEM and FEM codes: run them, understand their inputs/outputs, reproduce the rebounding sphere FEM validation, and trace how BEM fluid forces are passed to FEniCSx via preCICE.
Develop a working knowledge of airship fluid dynamics: added-mass effects, potential flow around a flexible envelope, and the physical consequences of fluid–structure coupling during a forced landing.
Be able to explain the full BEM–FEM–preCICE pipeline to a technical colleague: what each solver computes, what data crosses the interface, and which physical quantity is produced at each step.
Phase 2 – Implement, extend & demonstrate using LLM
Use Claude API to implement selected theoretical contributions from the Le Mestre thesis (operators, coupling terms, analysis methods not yet in the in-house codes) and validate them against thesis benchmarks.
Build the management demonstration tool: a simulation of the forced landing scenario showing quantitatively the effect of added masses on impact dynamics, with clear visual outputs comparing dry and wet cases.
Document the LLM methodology: where Claude accelerated development, where it introduced physically wrong results, and which prompting strategies worked best for this class of scientific computing problem.
Deliver an assessment of what remains out of reach for LLM-assisted scientific computing today.
YOUR EDUCATION & EXPERIENCES:
Final-year Master’s or engineering school student with a strong numerical methods background (FEM mandatory, BEM a plus)
Mechanical or aerospace engineering background; double degrees are a plus
YOUR SKILLS:
Good understanding of fluid mechanics: potential flow, pressure forces, fluid–structure interaction concepts
Structural mechanics: modal analysis, dynamic simulation, contact mechanics basics
Proficient in Python on Linux: NumPy, SciPy, Git, command-line tooling
Hands-on experience with Claude or another LLM for code generation
Autonomous and rigorous: able to judge whether an LLM output is physically correct, not just syntactically valid
APPLICATION PROCESS:
An interview with Thomas, Talent Acquisition Specialist.
An interview with Vaishali, Data Science Engineer.
PRACTICAL INFORMATIONS:
Duration: 4 to 6 months
Contract: Internship
Workplace: Suresnes – 92
Available position: As soon as possible
Department: Flying Works
Point of contact: www.flying-whales.com
Reference: STAGFSICFM-2026
