Founding Engineer (Multiphysics & Physics AI) at Gradient Dynamics

You will lead the development of a unified platform that bridges the gap between traditional computational fluid dynamics (CFD) and modern machine learning. By building high-performance multiphysics solvers from scratch and training advanced neural operators, you will empower engineers to develop complex products faster and more sustainably through differentiable simulation.

• Develop and optimize GPU-native multiphysics solvers using finite volume or finite element methods to maximize throughput on HPC systems.
• Design, train, and benchmark neural operator architectures (like FNO or DeepONet) against existing numerical techniques for engineering applications.
• Architect systems at the intersection of computational geometry and parallel architectures to underpin how the platform handles complex engineering designs.

• Holds a PhD or equivalent industry experience in computational physics, applied mathematics, or machine learning for scientific applications.
• Possesses a proven track record of writing production-quality scientific code and numerical methods from scratch, beyond just scripting in existing frameworks.
• Demonstrates deep expertise in GPU programming (CUDA, JAX, or XLA) and a strong grasp of governing equations like Navier-Stokes and energy equations.

• Rare opportunity to build a unified simulation-ML system from first principles, eliminating the traditional wall between CFD experts and AI researchers.
• Direct impact on a full-stack platform designed to solve the world’s hardest engineering problems across aerodynamics, thermal management, and multiphysics.
• Work at the bleeding edge of Scientific ML (SciML), utilizing JAX, CUDA, and neural operators to redefine how industrial designs are represented and tested.