Max Robotics — Global Service Robot Directory + US Certification

About

Kush Hari is a pioneering researcher at the intersection of surgical robotics, cloud robotics, and semantic scene understanding. His most impactful contribution is **Orbit-Surgical** (27 citations), a physics-based simulation framework that accelerates robot learning for surgical augmented dexterity, addressing a critical gap in fast and accurate surgical simulation. Hari also developed **Language-Embedded Gaussian Splats (LEGS)** (7 citations), enabling mobile robots to incrementally build room-scale, semantically rich 3D maps for object search in complex environments. His **Lifelong LERF** system (5 citations) advances inventory monitoring by allowing robots with limited compute to maintain dense language-geometric representations over time. In cloud robotics, Hari created **FogROS2-Config** (4 citations), a toolkit that helps roboticists select cost-effective cloud server configurations from over 50,000 options. His work on **FogROS2-FT** (2 citations) introduces fault tolerance for cloud robotics, ensuring reliability despite network or service disruptions. With over 60 total citations across his publications, Hari’s research spans surgical subtask learning, zero-shot policy transfer via **Mirage** (2 citations), and high-quality demonstration data generation for visuomotor policy learning. His contributions are driving progress in dexterous manipulation, cloud-robot reliability, and semantic mapping, making him a rising leader in robotics.