Sebastian Castro
Date:
Speaker
Sebastian Castro works on manipulation skill composition at the Robotics and AI Institute. He holds Bachelor’s and Master’s degrees from Cornell University in mechanical engineering, with a concentration on dynamics, systems, and control, applied to high-level planning and control of modular robots. His professional experience includes technical content development and marketing for robotics competitions with MathWorks, and software engineering at the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL), Boston Dynamics, and PickNik. Sebastian also devotes personal time to robotics education through blog posts, open-source software, mentorship, talks, and workshops. Speaker Links: Website
Abstract
In this talk, Sebastian outlines two dichotomies that keep roboticists awake at night: hand-engineered vs. machine learning based solutions, and approachable vs. unapproachable software design. While these seem unrelated at the outset, there are similarities in how designing systems to be understood by a broader audience can provide downstream benefits. Sebastian will present case studies inspired by his own experiences in task and motion planning, manipulation, and robot learning to argue for approachable systems design.