Machine Learning Engineer - Sim2Real & Machine Modeling
Gravis Robotics10 hours ago
Zürich, SwitzerlandSenior
Responsibilities
- Build machine-learning models to bridge the sim-to-real gap.
- Select and validate architectures such as sequence models and state-space formulations using data.
- Characterize unmodeled effects affecting sim-to-real transfer and determine which can be ignored safely.
- Determine data requirements and coverage distributions for modeling.
- Define performance metrics and validation methods for model fidelity and sim-to-real transfer.
- Build methods to detect changes in machine properties over time.
- Collaborate with autonomy and simulation teams to influence controllers running on machines.
- Collect missing data on site and interact directly with construction machinery when necessary.
Requirements
- Degree in Computer Science, Robotics, Machine Learning, Engineering, or a related field.
- Strong Python, PyTorch, and Git skills.
- Experience modeling time-series or dynamical-system data from large datasets using sequence models, system identification, or state-space approaches.
- Strong analytical skills, including experiment design, ablation studies, and defensible conclusions.
- Reinforcement learning experience is preferred.
- Imitation learning or learning from demonstration experience, especially with human operator data, is preferred.
- Familiarity with recent literature and methods in learned behavior policies is preferred.
- Classical system identification, control, or hydraulics experience is preferred.
Benefits
- Opportunity to join a dynamic, multidisciplinary Series A robotics company reshaping the construction industry.
- Work with real robotic systems and heavy construction machinery across laboratory and field environments.
- Equal opportunity employer committed to an inclusive and diverse workplace.
Categories
ML EngineeringRobotics
About Gravis Robotics
Gravis Robotics builds retrofit autonomy systems that turn heavy earthmoving machines into robots for construction contractors and equipment partners, combining LiDAR/camera/GNSS sensing with autonomous control and augmented teleoperation so one operator can supervise a fleet. Founded in 2022 as an ETH Zurich spin-out and headquartered in Zurich, it sells the Gravis Rack hardware with autonomy software and reports deployments in multiple countries; SoftBank is an investor.