Robotics Engineer – Imitation Learning & Autonomous Manipulation

About the role

FacilityOps.AI is seeking a Robotics Engineer to develop imitation‑learning pipelines and autonomous manipulation for robot arms that inspect and clean fiber‑optic connectors in data centers. You will work on both bench‑top and mobile robot platforms, turning human demonstrations into reliable, safe robot behaviours.

Key responsibilities

• Design and collect tele‑operated demonstration data using leader‑follower arms, VR or SpaceMouse interfaces.
• Train and evaluate behaviour‑cloning policies (e.g., ACT, Diffusion Policy) with PyTorch and integrate them with classical motion planning.
• Develop simulation environments in MuJoCo or Isaac Sim for data augmentation and sim‑to‑real transfer.
• Implement safety mechanisms such as force limits, no‑contact zones, and automatic failure recovery.
• Package policies as ROS 2 nodes, deploy on Kinova Gen3 or Hello Robot Stretch hardware, and support remote debugging.

Required profile

• Degree in robotics, mechatronics, mechanical, electrical, computer engineering, computer science, or equivalent hands‑on experience.
• At least 2 years of experience building robot manipulation systems, with a complete project from data collection to a working policy on a real arm.
• Proven ability to train and evaluate an imitation‑learning policy using PyTorch.

Required skills

• ROS 2 (nodes, topics, actions, TF, launch files) and MoveIt 2.
• Strong Python programming; working knowledge of C++.
• RGB‑D perception, object pose estimation, point‑cloud processing, OpenCV.
• Linux, Git, Docker, cloud GPU training.
• Simulation tools: MuJoCo or Isaac Sim; experience with sim‑to‑real transfer.
• Familiarity with Kinova, Hello Robot Stretch, LeRobot, or similar manipulators.