Max Robotics — Global Service Robot Directory + US Certification

About

Luca Weihs is a leading researcher in Embodied AI, focusing on building intelligent agents that perceive, navigate, and manipulate objects in simulated and real-world environments. His most impactful contribution is the **AllenAct framework** (44 citations), a foundational platform that standardizes training and evaluation for embodied agents, accelerating progress across computer vision, NLP, and robotics communities. Weihs also pioneered **ManipulaTHOR** (7 citations), enabling complex visual object manipulation tasks that go beyond simple navigation. His recent work tackles long-horizon manipulation with the **Universal Visual Decomposer** (8 citations), breaking down multi-stage tasks into learnable subtasks for improved policy generalization. Addressing real-world deployment, Weihs introduced **Promptable Behaviors** (6 citations), personalizing robotic behavior through multi-objective reward learning from human preferences, and **Seeing the Unseen** (3 citations), a novel visual common sense task requiring agents to reason about absent objects in scenes. His research consistently pushes toward disturbance-free, human-aligned robotic interaction, making him a pivotal figure in bridging simulation-based learning to practical, customizable embodied systems.