Robotics Engineer - Learning for Manipulation
Core
Train, fine-tune, and deploy learning-based manipulation policies using vision and force/torque inputs to enable robots to learn contact-rich manipulation.
Role type
Robotics engineer specializing in learning for manipulation
Builds
Learning-based manipulation policies and data-collection pipelines for real robot hardware
Domain
Robotics, Machine Learning, Computer Vision, Force Sensing
Deliverable
production ML models
Required skills
Python, C++, ROS 2, ML model training and fine-tuning, real robot hardware integration, software architecture design
Preferred skills
Robot learning frameworks (LeRobot), policy architectures (ACT, diffusion policy, VLAs), multi-modal/tactile learning, simulation (MuJoCo, Isaac Sim), force/impedance control
Technologies
ROS 2, Python, C++, MuJoCo, Isaac Sim, LeRobot, ACT, diffusion policy, VLAs
Responsibilities
Train and deploy manipulation policies using vision and force/torque inputs; build data-collection pipelines on real hardware; integrate physics-consistent simulation; design software architecture connecting sensors and robots; collaborate with hardware and embedded engineers
Seniority
Mid-level, hands-on IC
