CarbonSix says its toolkit brings robot imitation learning to the factory floor - The Robot Report

CarbonSix Inc., a manufacturing-focused AI developer, has launched SigmaKit, a toolkit that brings robot imitation learning directly to factory floors. Designed for industries such as electronics, batteries, and food and beverage, SigmaKit enables manufacturers to deploy AI-powered robots without needing specialized expertise or extensive programming. The system uses imitation learning, allowing robots to learn complex, delicate, and non-standardized tasks—like film attachment, assembly, and cable fastening—by observing human demonstrations. This approach addresses traditional automation challenges in unstructured and variable manufacturing environments, where conventional robots struggle to adapt. SigmaKit integrates AI algorithms, precision robotic grippers, sensor modules, and a teaching tool that simplifies operation. Users build training datasets by repeatedly demonstrating tasks, with models generated in less than a day depending on complexity. The resulting robotic "skills" can then be deployed to handle variable, non-routine tasks efficiently. Founded in 2024 by co-CEOs Jehyuk Kim and Terry Moon, CarbonSix leverages deep