Master Thesis Title: Physical Recursive Self Improvement (P-RSI)
Core
Develop a recursive self-improvement (RSI) system for physical robotics that automates the sim-to-real transfer loop and optimizes training algorithms and environments.
Role type
Master Thesis Researcher (Physical AI / Robotics)
Builds
An AI system capable of self-improving its own training pipeline for physical robots.
Domain
Robotics, Reinforcement Learning, Physical AI
Deliverable
research
Required skills
Reinforcement learning, Sim-to-real transfer, Robotics control, Automated code generation, Partial observability handling
Preferred skills
Automated unit testing, Dynamic environment modeling, High-cost failure mitigation
Technologies
Robotics stacks, Simulation environments, State-of-the-art computing infrastructures
Responsibilities
Design and implement RSI mechanisms for physical robots, Conduct sim-to-sim and sim-to-real transfer experiments, Validate and refine training rules across diverse scenarios, Minimize physical risk during real-world trials