Principal Architect, Reinforcement Learning (Physical AI)

What you'll do

• Design the end-to-end architecture for reinforcement learning and imitation learning pipelines to train physical AI models.
• Formulate tasks as observations, actions, and rewards while selecting appropriate policy and model architectures.
• Structure large-scale closed-loop training in simulation using GPU-parallel simulators and hybrid neural world models.
• Develop and measure sim-to-real transfer strategies using domain randomization, system identification, and actuator modeling.
• Design reward functions and metrics tied to productivity, safety, machine health, and operator-quality outcomes.
• Establish methodologies for policy evaluation, regression testing, and safety-case verification before field deployment.
• Set standards for experiment reproducibility, benchmark hygiene, and scalable training infrastructure.
• Translate research into repeatable engineering solutions for autonomous construction equipment.

What we're looking for

• Must be able to work full-time onsite at the Irving, TX office five days a week.
• Must possess knowledge of requirement analysis, analytical thinking, platform architecture, and target architecture.
• Must demonstrate effective communication skills to transmit and interpret ideas and business needs.
• Must be able to develop detailed architecture deliverables and design technical infrastructure for applications.
• Must lead the evaluation and deployment of new technologies to enhance digital technical capabilities.
• Master’s or doctoral degree in Computer Science, Robotics, Machine Learning, Controls, or a related field (preferred).
• Typically 8+ years (or PhD plus 5+) in robot learning with RL taken from algorithm to deployment on real hardware (preferred).
• Proficiency in Python, PyTorch, C++, CUDA, and experience with software engineering tools like Bazel or CI/CD (preferred).