Physics AI Machine Learning Engineer

Role Description

EMA3D is building the next generation of tools that put machine learning alongside numerical electromagnetic simulation. Our Physics AI team develops surrogate models that predict simulation results in seconds rather than hours or days and ships them directly inside our EMA3D Connect platform so engineers can use them in their everyday workflow.

We are looking for a Machine Learning Engineer to generate and curate simulation datasets, design and train the models, evaluate them and quantify their uncertainty, and build the pipelines that deploy them into production.

The team is very new, and strong direction is not always defined. You should be able to work in a fast-paced environment where the path to the solution is sometimes unknown.

Responsibilities

• Develop and train custom machine learning models that augment or accelerate electromagnetic numerical simulation (e.g., surrogate models, reduced-order models).
• Collaborate with physicists, engineers, and software engineers to design ML architectures suited for simulation data (grid-based, mesh-based, time-series, and multimodal scientific data).
• Implement pipelines for data preprocessing, feature extraction, and dataset management from large-scale simulation outputs.
• Evaluate model performance using physics-informed metrics (accuracy, stability, generalizability, interpretability).
• Deploy ML models into production workflows, including integration with simulation software and visualization tools.
• Stay current with advancements in scientific machine learning, physics-informed ML, and agentic AI systems, bringing innovative approaches into our products.

Qualifications

• Bachelor’s or Master’s degree in Computer Science, Electrical Engineering, Applied Mathematics, or related field (M.S. preferred for advanced ML/physics work).
• Strong proficiency in Python and ML frameworks such as PyTorch; C++ experience is a plus.
• Experience building and training ML models for scientific or engineering applications; familiarity with surrogate modeling, PINNs, or scientific neural operators, and graph neural networks is a plus.
• Knowledge of numerical simulation methods (FDTD, FEM, or other PDE solvers), ideally within electromagnetics or related physics domains.
• Experience working with large scientific datasets and high-performance computing environments.
• Strong problem-solving skills, ability to collaborate across multidisciplinary teams, and interest in bridging ML with domain science.