Scale-Resolved CFD Solver Developer - High-Order Methods, C++/CUDA and Turbomachinery
Role details
Job location
Tech stack
Job description
This is a full-time role for a senior CFD solver developer or computational scientist. You will join the R&D team developing our next-generation high-fidelity flow simulation software for turbomachinery. You will work directly on scale-resolved CFD numerics, high-order methods, solver architecture, GPU implementation and verification against demanding reference cases., * Extend PrediQT's scale-resolved CFD solver technology toward turbomachinery complex applications, such multi-stage, rotating stall or surge...
- Design, implement and optimize native GPU kernels and solver components in modern C++ and CUDA.
- Verify and validate new numerical developments on canonical CFD benchmarks, documented experiments and industrially relevant turbomachinery or propulsion cases.
- Contribute to scientific publications and represent PrediQT at leading CFD, aerospace, propulsion, turbomachinery, energy and weather-forecasting conferences.
- Write clear technical documentation, verification reports and, where appropriate, peer-reviewed scientific papers.
Requirements
We are looking for exceptional candidates who have already developed or contributed significantly to scale-resolved CFD solver technology and who can operate at the level of numerical methods, software architecture and GPU implementation., We are looking for candidates with a strong research and implementation background in CFD solver development. The ideal candidate has a PhD or equivalent research experience, has developed or substantially contributed to scale-resolved CFD codes, and is comfortable moving from numerical formulation to robust, high-performance C++/CUDA implementation.
This role is intended for candidates who have programmed the core numerical engine of CFD codes, not only used CFD packages or extended them through UDFs, scripts or isolated physical models. Relevant experience should include hands-on work on discretization schemes, fluxes, time integration, linear or nonlinear solvers, data structures, parallelization and performance-critical implementation.
Education details
A PhD in computational fluid dynamics, applied mathematics, mechanical or aerospace engineering, scientific computing, or a closely related field is strongly preferred. Exceptional candidates with equivalent research-level experience in CFD solver development may also be considered., * Deep hands-on knowledge of advanced CFD numerical methods at implementation level, including finite-volume and high-order approaches for compressible, turbulent or unsteady flows.
- Direct experience with high-order CFD methods such as Discontinuous Galerkin, spectral difference, flux reconstruction or related scale-resolved approaches.
- Expert-level modern C++ and strong CUDA programming skills; Python experience for tooling, testing and analysis is a plus.
- Strong expertise in scale-resolved CFD, including LES, DNS, hybrid RANS/LES or equivalent high-fidelity simulation approaches.
- Solid understanding of GPU architecture, memory hierarchy, parallel performance bottlenecks and optimization strategies for large-scale scientific computing.
Soft Skills
- Excellent analytical and problem-solving skills, intellectual curiosity, scientific rigor, clear communication and the ability to work effectively in a high-level R&D team.
- Creative, communicative and team working
- Experience in aerospace, propulsion, turbomachinery, energy, weather or climate applications is a strong advantage.
Preferred Skills
- Prior experience developing production or research CFD solvers, especially for turbomachinery, compressible flows, multiphysics or industrial high-performance simulations.
- Experience with CUDA is highly valued; experience with Kokkos, MPI, OpenMP, OpenACC or other HPC portability and parallelization frameworks is a plus.
Benefits & conditions
We offer a competitive compensation package, including salary, stock options and social benefits. This is an opportunity to join PrediQT at an early stage and contribute directly to the development of a new generation of high-fidelity industrial simulation software.