Max Robotics — Global Service Robot Directory + US Certification

About

Nikhil J. Joshi is a leading researcher at the intersection of robotics, natural language processing, and embodied AI. His work focuses on grounding large language models in real-world robotic affordances, enabling robots to understand and execute complex, temporally extended instructions. Joshi’s most influential paper, “Do As I Can, Not As I Say” (2022, over 500 citations), introduced a groundbreaking framework that bridges the semantic knowledge of LLMs with physical robot capabilities, addressing the critical gap between language understanding and actionable behavior. He also contributed to RT-1, a robotics transformer that scales real-world control by leveraging diverse, task-agnostic datasets for zero-shot and few-shot learning. More recently, Joshi led the development of RoboVQA (2024), a multimodal reasoning system that enables long-horizon task execution through a novel, bottom-up data collection scheme achieving 2.2x higher throughput than traditional methods. His work has been recognized for its practical impact on scalable, generalist robot learning, with applications ranging from household assistance to industrial automation. Joshi’s research continues to push the boundaries of how robots perceive, reason, and act in unstructured environments.