Max Robotics — Global Service Robot Directory + US Certification

About

Jarek Rettinghouse is a leading researcher at the intersection of robotics, natural language processing, and deep reinforcement learning. His work focuses on grounding high-level language understanding in physical robotic action, bridging the gap between semantic knowledge and real-world manipulation. Rettinghouse’s most influential contribution is the seminal paper “Do As I Can, Not As I Say: Grounding Language in Robotic Affordances” (2022), which has garnered over 516 citations. This work pioneered the use of large language models to encode actionable knowledge for robots, enabling them to follow complex, temporally extended instructions by constraining language generation to physically feasible actions. Beyond this breakthrough, Rettinghouse has demonstrated the practical scalability of deep RL through his system for sorting recyclables and trash in office buildings using a fleet of mobile manipulators (2023). This real-world deployment, though early in citation impact, showcases his commitment to translating algorithmic advances into tangible, socially beneficial applications. Rettinghouse’s research is defining how robots can understand and act upon human language in unstructured environments, making him a pivotal figure in embodied AI.