Job Description:

We are seeking a Robot Learning Engineer (Reinforcement Learning / Imitation Learning) to build learning-based systems that enhance autonomy, adaptability, and robot behavior.
You will develop RL/IL models that teach ROZOR robots how to navigate complex spaces, refine their cleaning or delivery behaviors, and respond intelligently to dynamic indoor environments.

Your work will directly shape how ROZOR robots learn, generalize, and operate with greater autonomy in real-world deployments.

Responsibilities:

• Develop RL and IL algorithms for navigation, manipulation, and task execution
• Build training pipelines in simulation for scalable robot learning
• Train models for dynamic obstacle avoidance, policy optimization, and adaptive behavior
• Use trajectory data, demonstrations, and sensor logs to create IL datasets
• Transfer learned policies to real-world ROZOR robots
• Build reward functions, curriculum strategies, and evaluation metrics
• Analyze simulation-to-reality (sim2real) gaps and improve transfer performance
• Collaborate with autonomy, perception, and embedded teams to integrate learned policies
• Conduct field validation and tuning of learning-based behaviors

Preferred Qualifications:

• Bachelor’s/Master’s in Robotics, AI, Computer Science, or related field
• Strong background in reinforcement learning, imitation learning, or decision-making algorithms
• Proficiency with PyTorch or TensorFlow
• Experience with simulation frameworks (Isaac Gym, Gazebo, Webots, Unity, Mujoco, etc.)
• Understanding of robot kinematics, navigation, and control systems
• Experience training RL/IL models and tuning reward functions
• Familiarity with ROS/ROS2 and real robot testing workflows
• Strong mathematical foundation in optimization, probability, and machine learning