Max Robotics — Global Service Robot Directory + US Certification

About

Mohan Kumar Srirama is a robotics researcher advancing the frontier of generalizable robot manipulation through large-scale data and representation learning. His work tackles the fundamental challenges that have long limited progress in robotics: the expense and diversity of hardware, and the lack of internet-scale training data. Srirama is a key contributor to the **DROID dataset** (108 citations), a landmark large-scale, in-the-wild robot manipulation dataset that provides diverse, high-quality data collected across multiple environments, enabling more robust and capable manipulation policies. He also introduced the **Train Offline, Test Online** benchmark, a real robot learning benchmark designed to democratize robotics research by allowing labs to train policies offline and evaluate them on standardized hardware. His work on **Human affordances for Robotic Pre-training (HRP)** explores how human interaction data can be leveraged to learn powerful visual representations for robots, reducing the need for robot-specific data. Through these contributions, Srirama is helping to build the data infrastructure and learning paradigms necessary for robots to operate reliably in the unstructured real world, making him a leading voice in the push toward general-purpose robotic manipulation.