Robotics Engineering Intern

Shepherd Robotics is building foundational robotic systems for cable and wire manipulation. We train robots to perform challenging real-world manipulation tasks involving flexible and deformable objects.

We’re looking for a Robotics Intern to help us collect high-quality robot data and improve our robot learning systems.

A large part of this role involves teleoperating real robots to generate training data. You’ll also have opportunities to contribute to model training, evaluations, data analysis, robotics infrastructure, and robot learning research based on your interests and strengths.

What you'll do

• Teleoperate robotic arms to perform cable and wire manipulation tasks
• Collect large amounts of high-quality demonstration data for training robot policies
• Learn difficult manipulation tasks and continuously improve speed, precision, and consistency
• Identify failure modes and help improve our task setups, controls, and data collection interfaces
• Evaluate trained robot policies and help diagnose where models fail
• Work with researchers on dataset quality, model training, and robot learning experiments
• Contribute across the robotics stack if you’re interested — software, hardware, controls, perception, or learning

We're looking for

• Excellent hand-eye coordination and the ability to learn physical tasks quickly
• Comfortable spending significant time operating and experimenting with real robots
• Strong attention to detail and consistency
• Excited about robotics and robot learning
• Prior robotics or teleoperation experience is a plus, but not required
• Competitive gamers and people with exceptional controller / mouse-and-keyboard skills are especially encouraged to apply
• Experience with VR controllers, RC systems, simulators, robotics competitions, or other high-dexterity interfaces is a major plus
• Python or ML experience is helpful but not required
• You don’t need to already be a robotics researcher. We’re especially interested in people who are extremely good at operating systems, learn quickly, and want to understand how modern robot learning actually works.
• Shepherd is YC-backed and supported by leading VCs and angels. We’re working on real deployment opportunities across automotive manufacturing, cable manufacturing, and hyperscale infrastructure.
• You’ll work directly alongside our robotics researchers and see the full loop from human demonstration → dataset → model training → evaluation → deployment on a real robot.