Robotics Engineering Intern | GoMyRobotARM (RL Training)
Company Description
GoMyRobot focuses on building physical intelligence infrastructure for space systems that can work together, operate autonomously, and remain serviceable over time. The company develops advanced robotic solutions that enable resilient, scalable, and efficient operations in space environments. Team members collaborate on cutting-edge technologies that integrate hardware, software, and automation. Applicants can expect to contribute to innovative projects that support the future of autonomous space systems while working with a technically driven, mission-oriented team.
Role Description
As a full-time remote Robotics Engineering Intern at GoMyRobot, you will support the training pipeline for GoMyRobotARM, our open-source manipulator arm for in-orbit servicing and autonomous rendezvous. Concretely, you will work within the BALERION stack: building MuJoCo simulation environments and integrating them with ROS 2 via mujoco_ros2_control and ros2_control's hardware-interface/lifecycle layer; training and tuning Soft Actor-Critic (SAC) policies alongside imitation-learning baselines (Diffusion Policy, ACT); and validating trained policies against the arm's real-time Control Barrier Function safety filter (500 Hz) prior to ONNX export and hot-swap deployment. Daily tasks may include running training experiments, profiling policy performance and sample efficiency, debugging sim-to-real discrepancies, documenting results, and working with the safety-layer team to ensure trained policies stay within CBF-constrained action bounds. You will also contribute to process automation initiatives and communicate progress and challenges with mentors and cross-functional team members.
Qualifications
- Foundational skills in robotics and a strong interest in autonomous systems, particularly manipulation
- Analytical skills to interpret training/evaluation data, diagnose policy failure modes, and reason about sim-to-real gaps
- Process automation skills to support experiment pipelines and automated training/evaluation workflows
- Communication skills to collaborate effectively in remote teams and document technical work clearly
- Software development skills in Python (PyTorch or similar) and/or C++, sufficient to modify and extend ROS 2 nodes
- Currently pursuing or recently completed a degree in Robotics, Mechanical Engineering, Electrical Engineering, Computer Science, or a related field
- Working knowledge of ROS 2 (rclpy/rclcpp), ros2_control, and MuJoCo; direct experience with mujoco_ros2_control is a strong plus
- Familiarity with RL training pipelines (SAC or comparable actor-critic methods) and, ideally, exposure to imitation learning (Diffusion Policy, ACT) or ONNX-based model deployment
- Ability to work independently, manage time in a remote environment, and learn new tools quickly
To apply: send your CV and portfolio to [email protected]