Physics-Informed ML Engineer: Model Architectures

1 week ago Be among the first 25 applicants

Field AIis transforming how robots interact with the real world. We are building risk-aware, reliable, and field-ready AI systems that address the most complex challenges in robotics, unlocking the full potential of embodied intelligence. We go beyond typical data-driven approaches or pure transformer-based architectures, and are charting a new course, with already-globally-deployed solutions delivering real-world results and rapidly improving models through real-field applications.

We are seeking a Physics-Informed Machine Learning (PIML) Engineer to join our innovative team focused on advancing risk-aware, autonomous systems. This role blends cutting-edge machine learning techniques with a strong foundation in physics, with an emphasis on safety, uncertainty quantification, and system robustness in real-world applications. The ideal candidate will work on integrating physical laws and constraints into machine learning models to create systems that learn from fewer data points while maintaining high accuracy and reliability in critical environments.

Field AIis transforming how robots interact with the real world. We are building risk-aware, reliable, and field-ready AI systems that address the most complex challenges in robotics, unlocking the full potential of embodied intelligence. We go beyond typical data-driven approaches or pure transformer-based architectures, and are charting a new course, with already-globally-deployed solutions delivering real-world results and rapidly improving models through real-field applications.

We are seeking a Physics-Informed Machine Learning (PIML) Engineer to join our innovative team focused on advancing risk-aware, autonomous systems. This role blends cutting-edge machine learning techniques with a strong foundation in physics, with an emphasis on safety, uncertainty quantification, and system robustness in real-world applications. The ideal candidate will work on integrating physical laws and constraints into machine learning models to create systems that learn from fewer data points while maintaining high accuracy and reliability in critical environments.

• Develop hybrid physics-ML models that combine theoretical physics-based components with data-driven elements to create more accurate and generalizable robotics autonomy solutions
• Design physics-informed architectures (e.g., physics-informed neural networks or universal differential equations) to solve complex robotic systems while respecting physical constraints like conservation of momentum, contact dynamics, and joint limits
• Lead research initiatives in physics-informed learning for robot control, combining model-based and model-free approaches, solving forward and inverse problems in robotic systems using PIML
• Create discrepancy models to bridge theoretical physics models with empirical data, analyzing the convergence, generalization, and error estimation of PIML models, ensuring stability and robustness in deployment
• Design and evaluate novel neural network architectures that respect physical laws and constraints
• Build and optimize differentiable simulation pipelines for robot trajectory and control policy optimization, addressing complex physical constraints such as uncertainty in perception systems
• Develop uncertainty-aware models combining physical knowledge with probabilistic state estimation (e.g., SDEs, Bayesian inference) for improved perception and intelligence
• Implement multi-scale modeling and domain decomposition to address large-scale challenges in autonomous robotics
• Collaborate with robotics teams to deploy physics-informed models in real-world autonomous systems
• Publish research in physics-informed machine learning and hybrid modeling for robotic systems

• Ph.D. or M.S. in Computer Science, Physics, Applied Mathematics, or related field with focus on robot learning and physical systems
• Track record of combining physics-informed machine learning techniques, with practical experience applying them to robotic systems
• Experience integrating physical constraints into machine learning architectures
• Strong understanding of POMDPs, differential equations, numerical methods, and computational physics
• Proficiency in implementing both physics-based and machine learning models
• Knowledge of conservation laws, symmetries, in variances, and conservation laws relevant to robotic systems (e.g., SE(3) equivariance, Lie groups, Noether’s theorem to encode symmetries and in variances into geometric deep learning models for robotics)
• Experience with differentiable programming frameworks (PyTorch, JAX) and robotics middleware
• Strong programming skills in Python, C++, or Julia, with experience deploying algorithms on real robots
  

Compensation and Benefits

Our salary range is generous ($70,000 - $200,000 annual), but we take into consideration an individual's…