deine aufgaben
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 RobCo’s 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
dein profil
Degree in Machine Learning, Robotics, Computer Science, Mathematics, Engineering, or a related field (Master’s 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
warum wir?
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