I’m Kyle Vedder

I am a researcher at Physcial Intelligence, where I work on building self-improving robotics systems. Before that I worked at Dyna Robotics, where I did full stack robot learning, touching every part of the data collection, model training, and inference stack.

I did my CS PhD at Penn, where I was advised by Eric Eaton in the GRASP Lab. My PhD research focused on scene flow: building better evaluations, and developing a repertoire of tricks to make scalable, flexible scene flow a reality. I was fortunate to work closely with Deva Ramanan, James Hays, and Dinesh Jayaraman throughout my PhD. For a short narrative overview of how my PhD research fits together, see Overview of my PhD Research. For a longer overview with nice pictures, check out the intro of my thesis.