Machine Learning Engineer
Root Access2 months ago
New York, NY, USAMid Level
Responsibilities
- Design and train physics foundation models using PINNs, Fourier Neural Operators, and Neural Operators to solve Maxwell’s, Helmholtz, and heat equations within neural loss functions.
- Build ECAD data pipelines that convert ODB++, IPC-2581, STEP, and Gerber PCB files into tensor grids, signed distance fields, or graph embeddings.
- Implement differentiable physics calibration pipelines using VNA Touchstone files, TDR traces, and near-field EMI scans to tune material and manufacturing parameters.
- Integrate upstream Graph Neural Networks or LLMs with downstream spatial physics engines.
- Optimize GPU-cluster training and inference pipelines for sub-100-millisecond physics predictions and real-time layout-design feedback.
Requirements
- Master’s or Ph.D. in Computer Science, Mathematics, Electrical Engineering, Physics, or a related quantitative field, focused on Scientific Machine Learning.
- At least four years of expert-level experience with PyTorch or JAX.
- Hands-on experience building and training PINNs, DeepONets, or Fourier Neural Operators.
- Direct experience with NVIDIA Modulus, DeepXDE, or PyTorch Geometric.
- Exceptional understanding of partial differential equations, vector calculus, automatic differentiation, and numerical optimization algorithms including Adam and L-BFGS.
- Strong proficiency manipulating spatial or geometric datasets with Python libraries such as NumPy, SciPy, Shapely, Open3D, or custom voxelization matrices.