Tasks:
- * Build generalized reward and feedback models
- * Deploy models on real robots using ROS2
- * Develop post-training workflows from SFT to RL
- * Implement real-robot online RL and data collection loops
- * Improve VLA/WAM policies with reinforcement learning
- * Scale asynchronous training across multiple robots
- * Validate policies in simulation and sim-to-real
Perks/Benefits:
Skills/Tech stack required:
[Action Chunking] [Action Chunking with Transformers (ACT)] [CQL] [Curriculum learning] [Diffusion Policy] [Distributed Training] [Flow matching] [IQL] [Isaac Lab] [Linux] [Mixed Precision] [Mixed-precision training] [PPO] [Python] [Reinforcement Learning] [Residual Reinforcement Learning] [Reward shaping] [Robot Kinematics] [ROS2] [SAC] [Sim-to-Real] [TD3] [Teleoperation] [VLA] [World Models]
Educational requirements:
[Bachelor's Degree] [Master's Degree] [PhD]
Role(s):
[Algorithm Engineer] [Engineer] [Learning Engineer] [Reinforcement Learning Engineer] [Robotics Algorithm Engineer] [Robot Learning Engineer]