RoboWeave: AI-Generated Robot Code into Executable Missions

This project proposes a cloud-based fleet management platform that registers robots, catalogs their capabilities and operating contexts, and uses those records to generate robot-specific task scripts. An operator will describe a task in natural language. The platform will match the task to a compatible robot, supply an AI code generator with approved robot and physical-AI model APIs, and produce a Python execution script. After validation, operators can download the script to the target robot. A web-based teleoperation interface will let operators monitor execution, intervene when needed, and collect demonstrations for future task development.

Programming a robot for a new task often requires a developer to understand its hardware, available models, software interfaces, and operating environment. This becomes harder across a fleet whose robots have different capabilities. The project applies vibe coding to robotics: an operator describes the intended outcome, while an AI code generator assembles a task script from predefined APIs. Unlike general-purpose code generation, the proposed system grounds each script in a registered robot’s features and context, so it uses only capabilities the selected robot supports.

Team1 (Cloud Fleet Management)

• Build a cloud robot registry that records each robot’s hardware, software, capabilities, location or operating context, State, and deployment status. (Azure IoT Hub)
• Build an API registry for physical-AI models and robot functions, including their inputs, outputs, dependencies, constraints, and usage examples.
• Integrate Training data Collection for a Task that can be matched to a compatible Robot
• Develop a teleoperation dashboard to view task progress, intervene remotely, and record operator demonstrations.

Team 2 (Vibe Coding for task execution)

• Match tasks to robots based on required capabilities, model compatibility, and operating context.
• Generate and validate Python task scripts using only approved APIs available to the selected robot.
• Execute the generated task scripts on the target Robot Edge (NVIDIA Jetson Orin NX)
• (Optional) Develop Trained Models for various Sub-Tasks and validate.

Python, Cloud Programming/Automation (Azure) Experience, Front-End/Back-End Development

Azure IoT Hub, Robotics Experience, Claude Experience (or similar)