Date: October 22, 2025
Speaker: Carmelo (Carlo) Sferrazza, incoming Assistant Professor at UT Austin
Title: Humanoid robot learning
Abstract: Humanoid robots represent the ideal physical embodiment to assist us in the diversity of our daily tasks and human-centric environments. Driven by substantial hardware advancements, progress in artificial intelligence (AI), and a growing demand for adaptable automation, this vision appears increasingly feasible.
Yet, humanoid intelligence remains far from achieving the envisioned general-purpose capabilities. In this talk, I will discuss the unique challenges humanoids pose in the robot learning setting and present approaches to scale learning through novel tools (HumanoidBench, MuJoCo Playground), flexible algorithms (OmniRetarget, FastTD3), and expressive architectures (Body Transformer).
Bio: Carmelo (Carlo) Sferrazza is an incoming Assistant Professor at UT Austin. He is currently with Amazon Frontier AI & Robotics (FAR), and until recently was a postdoctoral researcher at UC Berkeley with Prof. Pieter Abbeel. His research focuses on advancing humanoid robots’ intelligence and loco-manipulation capabilities by incorporating priors, inductive biases, and multi-sensory feedback. He earned his Ph.D. from ETH Zurich under the supervision of Prof. Raffaello D’Andrea, where he developed vision-based, data-driven tactile sensors and explored their applications in robot control and dexterous manipulation. Carlo is a recipient of the Best Demo Paper Award at RSS 2025, the 2022 ETH Medal, the Best Paper Award at RoboSoft 2022, and the 2017 ETEL Award. He was also selected as a Rising Star at RoboSoft 2025 and as an RSS Pioneer in 2022.