Max Robotics — Global Service Robot Directory + US Certification

About

Yunshuang Li is a robotics researcher whose work is driving progress in robot manipulation, particularly for long-horizon and real-world tasks. Her key research areas include large-scale robot learning, multimodal perception, and bio-inspired control. Li’s most significant contribution is the DROID dataset, a large-scale, in-the-wild robot manipulation dataset that has already garnered over 108 citations since its 2024 release. This dataset addresses the critical challenge of collecting diverse, high-quality manipulation data across varied environments, providing a foundational resource for training more robust and generalizable robotic policies. She also developed the Universal Visual Decomposer, which simplifies long-horizon manipulation by automatically breaking complex tasks into manageable subtasks, facilitating policy learning and generalization. Additionally, Li has contributed to the PEg TRAnsfer Workflow Recognition Challenge, investigating whether multimodal data improves task recognition. Her early work includes controlling pneumatic artificial muscles using a spiking neural network-based cerebellar model, showcasing her versatility. Through these efforts, Li is helping to bridge the gap between controlled lab settings and the unstructured, multi-stage tasks that robots must master for real-world deployment.