​A deep-tech startup developing next-generation nuclear fusion simulation technology is looking for Chinese-speaking engineers and scientists with strong backgrounds in simulation, scientific computing, numerical methods, computational physics, HPC or related fields.

What you’ll be working on:

Depending on your background and area of expertise, your work may include:

• Developing and optimizing numerical algorithms for large-scale scientific and engineering simulations, including numerical discretization, linear/nonlinear solvers and time-integration methods.

• Developing and optimizing CPU/GPU and large-scale parallel computing solutions, including MPI, CUDA/HIP, domain decomposition, communication optimization and load balancing.

• Developing computational models and simulation algorithms for plasma physics, fluid dynamics, transport, turbulence, electromagnetics and other complex physical systems.

• Engineering and optimizing scientific simulation software using C/C++/Fortran, including core algorithms, software architecture, visualization and simulation workflows.

• Improving simulation performance, accuracy, stability and scalability across different computing architectures.

• Building, testing and validating simulation models and software against benchmark cases and real-world applications.

• Working with researchers and engineers to translate mathematical models and algorithms into robust, scalable software.

• Keeping up with developments in scientific computing, numerical simulation, HPC and computational science, and applying relevant techniques to the platform.

Candidate Profile

• Open to candidates from computational physics, scientific computing, numerical methods, CFD/CAE, HPC, parallel computing, GPU computing, scientific software, plasma physics, applied mathematics and related fields.

• Relevant experience may include C/C++, Fortran, Python, CUDA/HIP, MPI, PETSc, hypre, Kokkos, RAJA, MATLAB, CMake or Linux.

• Experience in energy-related computational science, engineering simulation or high-performance computing can be highly transferable.