Software Engineer, Machine Learning in Applied Physics

Join nTop to apply for the Senior Software Engineer, Machine Learning in Applied Physics role. Engineering teams face an impossible reality : delivering more complex products faster, with fewer experts, and zero tolerance for failure. nTop changes how engineering gets done. Our technology collapses months of iteration into hours, letting teams explore thousands of variants instead of settling for the first option. Teams reduce development time by 50% and increase program win rates. Leaders choose nTop when failure isn't an option. If you're motivated by solving tough engineering challenges alongside a team that learns and grows together, you'll thrive at nTop. This role is remote in either Germany or the UK, and reports to the Engineering Manager, Analyze.

The role focuses on building cutting-edge surrogate models that accelerate complex physics simulations by 100-1000x, enabling real-time engineering design and optimisation.

What You'll Do

• Build and deploy surrogate models to replace expensive physics simulations with 100-1000x speedups
• Integrate existing ML solutions into the nTop ecosystem
• Prepare and curate training data by labeling simulation outputs, annotating physics features, and building high-quality datasets from multi-fidelity sources
• Validate models rigorously against high-fidelity simulations and experimental data with uncertainty quantification
• Document model architectures, assumptions, and limitations for technical and non-technical stakeholders

• MS / PhD in Physics, Applied Mathematics, Mechanical / Aerospace Engineering, or related computational field

• 3+ years developing surrogate / reduced-order models for physical systems

• Proven expertise in deep learning frameworks (PyTorch or TensorFlow) applied to scientific problems

• Strong foundation in numerical methods, PDEs, and computational physics (FEM, CFD, or similar)

• Proficiency in Python scientific computing stack (NumPy, SciPy, Pandas, scikit-learn)

• Experience with uncertainty quantification and Bayesian methods in ML models

• Track record of working with large-scale simulation data and HPC environments

• Hands-on experience with at least one physics simulation package (ANSYS, COMSOL, OpenFOAM, etc.) Preferred Experience

• Experience with neural operators (DeepONet, Fourier Neural Operators) for PDE solving

• Experience with NVIDIA ecosystem for physics AI (Physics Nemo, Omniverse, SimNet)

• Familiarity with NVIDIA Domino for ML model management and deployment Experience with setting up and deploying inference services

• Proficient in containerization technologies, particularly Docker

• Background in aerospace and energy systems

• GPU programming experience (CUDA, OpenCL)

• Understanding of digital twin architectures and real-time simulation requirements