Codex Pianist: LLM-Assisted Robotic Piano System
A robotic pianist controlling an Inspire robotic hand mounted on a UF850 robot arm, built using LLM-assisted engineering workflows. Top 5 Finalist at the OpenAI Codex Hackathon.
Top 5 Finalist at the OpenAI Codex Hackathon. Demonstrated how LLM-assisted engineering workflows can assemble complex hardware-software systems.
Built a robotic pianist controlling an Inspire robotic hand mounted on a UF850 robot arm. The system uses a Python control pipeline integrating robotics SDKs and Modbus device communication.
Approach
The project demonstrates how LLM-assisted engineering workflows can rapidly assemble complex hardware-software systems. Using OpenAI Codex, the development process was accelerated by leveraging AI to generate and refine control code for the robotic hand's finger articulation and the robot arm's positioning.
Technical Stack
- Python control pipeline with robotics SDK integration
- Modbus device communication for hardware control
- NVIDIA Isaac Sim for simulation and testing
- UF850 Robot Arm with Inspire Robotic Hand
- OpenAI Codex for LLM-assisted development
Key Achievement
Selected as a Top 5 Finalist at the OpenAI Codex Hackathon, demonstrating the practical application of LLM-assisted workflows in complex robotics integration.