Robot Learning Engineer (m/f/d) VLA & Foundation Models

The AI Research Division of Agile Robots is looking for a Robot Learning Engineer (m/f/d) VLA & Foundation Models, who will develop and deploy vision- and language-conditioned robot policies for real-world manipulation on physical robot platforms.

• Policy Development: Design, train, and evaluate vision- and language-conditioned robot policies using approaches such as imitation learning, diffusion models, transformer architectures, or reinforcement learning.
• Data Pipelines: Build workflows for collecting, structuring, and validating multimodal robot data from teleoperation, simulation, and real-world experiments, covering video, proprioception, force signals, and task outcomes.
• Robot Integration: Define action spaces, control interfaces, temporal chunking, and evaluation protocols to ensure learned policies are compatible with real robot execution contexts.
• Failure Analysis: Test policies in simulation and on physical robots, diagnosing failures caused by distribution shift, perception errors, contact dynamics, latency, and data quality.
• Deployment: Integrate learned policies into robot software stacks, with attention to latency, synchronization, safety checks, fallback behavior, and runtime monitoring.
• Research Application: Track developments in robot learning, VLA models, and foundation model methods, applying relevant advances to Agile's robot platforms and use cases.

• Background: Degree in computer science, robotics, electrical engineering, or a closely related field, or an equivalent combination of research and professional experience in robot learning or embodied AI.
• Robot Learning: Hands-on experience building or working with learning-based robot policies, including imitation learning, visuomotor control, VLA models, or closely related embodied AI methods.
• Robotics Systems: Practical experience with robotic hardware, including real robot manipulation, ROS/ROS2, sensor integration, control interfaces, or real-world robot deployment.
• ML Engineering: Working proficiency in Python and PyTorch, including the ability to implement models, training loops, data loaders, and evaluation pipelines beyond notebook-level prototypes.
• Experimental Practice: Ability to design experiments, compare model variants, debug failures across the ML/robotics boundary, and iterate toward measurable performance on real tasks.

• VLA Experience: Familiarity with VLA models or robotic foundation models such as π0, GR00T, or equivalent, including action-token or action-chunk representations.
• Policy Deployment: Experience deploying learned models on physical robots or edge compute platforms, including latency-constrained inference pipelines.
• Reinforcement Learning: Experience with online or offline RL for robotics, including reward design, simulation-based training, or combining RL with imitation learning.
• Simulation: Experience with simulators such as MuJoCo, Isaac Sim, or ManiSkill and with sim-to-real transfer workflows.
• Data Collection: Experience building or operating teleoperation systems, human-in-the-loop data collection, or intervention learning setups.

• 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.