Machine Learning Engineer – Robot Learning (mfd)

As a Machine Learning Engineer Robot Learning you will help develop evaluate and deploy learning-based methods for real-world robotic manipulation. You will work at the intersection of machine learning robotics and systems engineering adapting state-of-the-art research into robust scalable capabilities that run on real hardware.

You will collaborate closely with robotics autonomy perception simulation and software teams and you will own ML workflows from model design to deployment and testing on real robots.

Your Responsibilities

• Research evaluate and benchmark state-of-the-art robot learning methods (VLA models diffusion policies RL imitation learning visuomotor models)

• Adapt academic models for practical real-world deployment on RobCos modular robots

• Train and fine-tune ML models using RobCo datasets and simulation data

• Integrate learned policies with perception control and robot runtime systems

• Build scalable training evaluation and data pipelines in collaboration with infrastructure teams

• Define clear performance metrics and build automated evaluation procedures in simulation and on real hardware

• Analyze model performance identify regressions and drive improvements

• Collaborate with robotics and autonomy engineers to ensure safety reliability and real-time performance

• Participate in research planning and contribute to technical decisions across the robot learning stack

• Share knowledge support junior team members and help shape internal ML best practices

• Degree in Machine Learning Robotics Computer Science Mathematics Engineering or a related field (Masters degree necessary)

• Hands-on experience with robot learning imitation learning reinforcement learning or deep learning

• Strong coding skills in Python and experience with PyTorch or JAX

• Experience training or evaluating ML models (university projects internships research labs industry experience all count)

• Familiarity with robotics concepts or experience deploying ML on real systems (internship or lab experience is sufficient)

• Ability to design experiments analyze model behavior and communicate insights clearly

• Strong problem-solving mindset and eagerness to work on real robots

• Bonus: familiarity with ROS 2 scalable training tools (Ray Anyscale) or robotic datasets

• Shape the robot learning capabilities of a next-generation modular robotics platform

• Work with real hardware simulation tools rich datasets and scalable ML infrastructure

• Research-focused culture with direct real-world deployment

• High ownership autonomy and strong technical growth opportunities

• Hybrid work model flexible hours and cutting-edge equipment