Founding Engineer - Multiphysics Simulation & Physics AI

ABOUT GRADIENT DYNAMICSGradient Dynamics is a VC backed deep tech company on a mission to build the full stack Physics AI platform for engineering. We combine deep expertise in computational physics, scientific machine learning and HPC to solve some of the world’s hardest engineering problems. We aim to help companies develop better products, faster. We are committed to progressing physics AI to help advance engineering, creating a better, more sustainable future.We are currently validating our GPU-native simulation platform with early commercial partners and beta users across industry and academia. This is an opportunity to join at an early stage and help define the technical foundations of a category-defining company at the intersection of simulation, AI, and high-performance computing.THE ROLEWe are looking for a Founding Engineer to act as a core member of the technical team and help build the next generation of Physics AI infrastructure and applications.This is not a narrowly scoped research role. You will work across the full technical stack - from computational geometry and mesh generation to GPU-native solver development, dataset generation, scientific ML training, optimisation systems, and deployment infrastructure.If you've been frustrated by the artificial wall between traditional simulation (CFD, CHT, FEA) and modern ML - where the simulation people don't trust the ML people and vice versa - this is a chance to build the unified system from first principles. This role exists at the intersection of traditional simulation, GPU computing, and scientific machine learning. You should be as comfortable debugging a pressure-velocity coupling scheme as you are fine-tuning a neural operator on point-cloud data. If you have spent your career so far frustrated by the artificial separation between simulation and AI, this is your chance to build the unified platform from the ground up.This role is designed for engineers who want deep technical ownership, significant autonomy, and the opportunity to shape both product direction and core architecture from an early stage. You will work directly with the founders on technically difficult, high-impact problems with immediate commercial relevance.WHAT YOU WILL DOSolver development: Continue the development of our multiphysics solvers. Write high quality code and develop numerical methods.Physics AI: Train and evaluate neural operator models. Run fine-tuning experiments, benchmark against existing techniques, and iterate on model architectures.Performance Optimisation: Engineer complex codebases to maximise throughput on HPC systems.Geometry/Meshing: Work at the intersection of computational geometry, numerical methods, GPU/parallel architectures to help design the systems that underpin how our technology represents designs.Validation & Benchmarking: Develop validation methodologies, benchmarking pipelines, and reproducibility frameworks against analytical, experimental, and industrial reference data.WHAT SUCCESS LOOKS LIKEWithin your first 6-12 months, you will have;Contributed production-grade functionality to our multiphysics simulation stack and helped improve solver performance and scalability on GPU systems.Built or trained Physics AI models on proprietary simulation datasets and contributed to commercially relevant engineering optimisation workflows while shaping the long-term technical architecture of the platform.REQUIREMENTSEssential:PhD or equivalent industry experience in computational physics, applied maths, mechanical/aerospace engineering, or ML for science.Demonstrated ability to write production-quality scientific code from scratch - finite volume or finite element methods, not just scripting around existing frameworks.Strong Python and comfort with GPU programming concepts (CUDA, JAX XLA, or equivalent).Working knowledge of fluid dynamics and heat transfer at a governing-equations level (Navier-Stokes, energy equation, turbulence closures).Familiarity with scientific ML: neural operators (DeepONet/FNO), PINNs, differentiable simulation etc.Ability to work autonomously at startup pace, context-switching between solver numerics, data pipelines, and model training within the same week.DesirableDirect experience with Pytorch or JAX, or a track record of rapid framework adoption.Prior work on aerodynamics, thermal management, or multi-physics simulation.Published research in computational methods, scientific ML, or a related domain.Experience with cloud-based HPC workflows (AWS, GCP) and containerised simulation pipelines.WHY JOIN GRADIENT DYNAMICSWork on Fundamental Technical Problems: You will help build core infrastructure at the convergence of simulation, AI, geometry, and HPC — an area increasingly viewed as foundational to the future of engineering software.High Ownership & Technical Influence: As an early technical hire, your work will directly shape architecture, product direction, and technical strategy.Rare Technical Breadth: Few companies operate simultaneously across computational physics, GPU-native HPC, scientific ML, computational geometry and differentiable optimisation.Join Early: This is an opportunity to join before major scale-up, with meaningful ownership and the ability to influence company trajectory.BENEFITSCompetitive salary and meaningful equity participation.Work directly with founders and early technical leadership with opportunity to publish and contribute to cutting-edge technical work.Fast-moving, technically ambitious environment with significant autonomy and creative freedom.Opportunity to help define an entirely new category of engineering software.