RAPID: Robot Agentic Programming from Demonstrations

Yuyao Liu, Jiayuan Mao, David Hsu et al.1 min readarXivSep 24, 2026

Key Takeaway for Students

Coding agents have demonstrated enormous success in solving complex programming problems. To leverage their potential for robot systems, this work introduces Robot Agentic Programming from Demonstrations (RAPID), which automatically generates, verifies, and refines robot programs, given a single visual human demonstration. The iterative agentic loop of code refinement requires

Reading in Plain English mode — toggle below to switch

This paper introduces a novel approach for robot programming using agentic loops and visual demonstrations, significantly reducing the manual effort in developing robot behaviors. For ML engineers, understanding how to integrate LLMs with robotic systems for autonomous code generation and verification is crucial for advancing automation and intelligent systems.

Read original on arXiv

Career Insight

high demand

Why it matters

Skills to learn

Robot Operating System (ROS) fundamentalsVisual perception for robotics (e.g., object detection, pose estimation)LLM prompt engineering for code generationDesigning and implementing agentic loops for iterative refinementRelational program representation for robotic tasks

Interview questions

Describe the key components of RAPID's iterative agentic loop and how each contributes to program refinement.
How would you adapt RAPID's approach to a scenario where visual demonstrations are ambiguous or incomplete?
Design a system architecture for deploying RAPID in a manufacturing plant, considering real-time constraints, safety protocols, and integration with existing robotic hardware.

Relevant for ~24 months · generated by the enrichment pipeline