Intern - Reinforcement Learning Engineer

Description

Ghost Robotics is the industry leader in legged robotic systems that not only help our customers solve complex operational, national security, and technical challenges to save lives, reduce harm and improve outcomes.

We are looking for a part-time Reinforcement Learning Intern to work with the Chief Science Officer to develop reinforcement learning algorithms for quadrupedal robot self-righting.

Key Duties:

• Develop infrastructure - scripts for training and evaluating algorithm progress
• Use mjlab for training in simulation, with domain randomization for zero-shot sim2sim and sim2real transfer
• Hands-on testing on the physical robot
• Reward tuning for desirable performance

Requirements

Recommended Qualifications:

• Holds Bachelor's degree in Computer Engineering, Software Engineering, or a related field
• Currently pursuing Masters degree in Computer Engineering, Software Engineering, or a related field
• Expertise in reinforcement learning algorithms like PPO, TRPO
• Experience with weights & biases dashboard setup
• Python programming experience
• Mjlab or Isaac sim experience
• Optional but highly desirable: taken ESE 6510 Physical Intelligence course at Penn (or equivalent)

Location

Philadelphia, PA (no remote candidates considered at this time).

Schedule and Committment

Flexible, part-time schedule, 20 - 30 hours per week. Able to commit to three (3) month minimum project, with option to extend following evaluation.

Travel

No Travel Required.

Compensation

Competitive base $30 - $45 per hour.

Background Check

Clear standard background checks, pre-hire, post hire and anytime during employment as required.

Residency Requirements

Employment Authorization Required.

Intellectual Property

Note: An IP agreement needs to be signed for this project, and any public release of project outputs requires prior company approval.

Physical Requirements

• Prolonged periods of standing, sitting at a desk and working on a computer.
• Must be able to lift 20 pounds. Assistive equipment available.