(Senior) Reinforcement Learning Engineer (m/f/d) Manipulation & Locomotion

The AI Research & Development Group of Agile Robots is looking for a (Senior) Reinforcement Learning Engineer (m/f/d) Manipulation & Locomotion, who will apply reinforcement learning to existing Vision-Language-Action (VLA) and control models, making them reliable for dexterous manipulation, whole-body control, and humanoid walking on real hardware. The models themselves are built elsewhere in the organization; this role begins with an already-trained policy and focuses on making it perform reliably on the target robot and task.

• Policy Refinement: Apply reinforcement learning to an already-trained Vision-Language-Action (VLA) or other manipulation/control model, improving its reliability on a specific target task.
• Dexterous Manipulation: Own RL-based reliability work for contact-rich manipulation tasks, including grasping and fine object interaction.
• Whole-Body & Locomotion: Apply reinforcement learning to whole-body control and humanoid walking, improving balance and coordinated movement under real physical constraints.
• Reward & Task Design: Define the reward structures and task specifications that determine what counts as reliable performance for each policy-refinement effort.
• Sim-to-Real Transfer: Train and validate policies in simulation, then adapt them to close the performance gap on the physical robot.
• Failure Diagnosis: Identify where a refined policy breaks down in practice, and adjust reward design, task structure, or training approach in response.

• Background: Master's degree in robotics, computer science, machine learning, or a related field, or an equivalent combination of formal training and professional experience.
• Professional Experience: 2 to 5+ years of professional experience in robotics or robot learning engineering.
• Reinforcement Learning: Working understanding of how to fine-tune or refine an already-trained Vision-Language-Action (VLA) or other manipulation/control policy through reinforcement learning.
• Robotics Grounding: Hands-on experience with robot manipulation or legged locomotion on real, physical hardware.
• Sim-to-Real Transfer: Experience training policies in simulation and transferring them to physical robots, including a working understanding of where that transfer typically fails.
• Physical Interaction: Solid understanding of contact dynamics and physical constraints relevant to manipulation or legged locomotion.

• Broader Model Exposure: Familiarity with policy architectures beyond VLA, such as diffusion-based or transformer-based control models.
• RL Tooling: Familiarity with reinforcement learning libraries and simulation environments used for robot learning, such as Isaac Lab or Stable Baselines.
• Research Contribution: Contribution to peer-reviewed robotics or machine-learning research, for example at CoRL, ICRA, or RSS.
• Humanoid Deployment: Experience deploying learned policies on a bipedal or high-degree-of-freedom robot platform.

• Dynamic high-tech company combined with financial soundness and world class investors.
• Join an interdisciplinary, international team with 60+ different nationalities in a collaborative work environment.
• Lots of development opportunities in the context of our continued growth.
• Challenging tasks and impactful projects alongside experts that enable professional and personal growth.
• Corporate Benefits Program that covers health, mobility and learning with 100 € net per month.
• Modern office facilites with a rooftop terrace overlooking Munich, free drinks & fruits, and regular company events contribute to a good working environment.