Senior Machine Learning Engineer Reinforcement Learning World Model

Path Robotics develops physical-AI robotic welding systems for manufacturing.

Columbus, United States
About Path Robotics

Path Robotics builds autonomous welding cells and mobile welding systems powered by its Obsidian AI model, which uses real-time sensing and adaptation to weld variable industrial parts.

View jobs by Path Robotics

Skills

Candidate Availability

Required and preferred rules are kept separate and reflect the wording in the original posting.

About the Role

You will develop learned world models and reinforcement-learning systems for welding robotics. You will model visual and physical welding dynamics, train and evaluate policies, quantify uncertainty, validate predictions with real-world data, diagnose unsafe or unstable behavior, and turn research work into dependable training, evaluation, inference, and deployment systems.

Requirements

  • Master’s or PhD in Computer Science, Robotics, Machine Learning, or a related field, or equivalent practical experience.
  • Experience developing and deploying reinforcement-learning algorithms on real-world systems.
  • Proficiency in Python and deep-learning frameworks such as PyTorch or TensorFlow.
  • Experience with simulation environments such as MuJoCo or Isaac Gym.
  • Understanding of probability, statistics, and optimization.
  • Experience training and deploying machine-learning models in production systems.

Responsibilities

  • Build action-conditioned world models for welding-process dynamics.
  • Model inputs, system state, physical dynamics, and weld quality.
  • Develop multimodal models using video, 3D scans, thermal measurements, electrical signals, robot state, and process parameters.
  • Improve rollout accuracy, physical plausibility, temporal consistency, and computational efficiency.
  • Quantify model uncertainty and validate predictions against real-world welding data.
  • Integrate learned models into reinforcement learning, planning, process optimization, evaluation, and synthetic-data workflows.
  • Develop reinforcement-learning approaches for optimizing welding decisions and outcomes.
  • Define state, observation, action, and reward representations.
  • Train and evaluate policies using learned models, simulation, offline data, and controlled real-world experiments.
  • Diagnose reward exploitation, unsafe behavior, policy instability, and model exploitation.
  • Translate research prototypes into dependable training, evaluation, and deployment systems.

Benefits

  • Daily free lunch
  • Flexible PTO
  • Medical, dental, and vision coverage
  • 6 weeks of fully paid parental leave, plus an additional 6–8 weeks for birthing parents
  • 401(k) retirement plan through Empower
  • Employee referral bonuses
Senior Machine Learning Engineer Reinforcement Learning World Model at Path Robotics | JobStash