LeRobot Dataset v3.0: Scaling Reads with Ray | Anyscale | Ray Summit 2026
LeRobot Dataset v3.0 is an emerging standard for robot manipulation data, but scaling reads from it is challenging: the format splits each episode across multiple video files while concatenating many episodes into the same file, so naive parallel processing requires opening each MP4 many times. At Ray Summit 2026, Artur Niederfahrenhorst, Member of Technical Staff at Anyscale, looks at the format and at approaches for using it efficiently in Ray-powered pipelines. You'll leave knowing how to read LeRobot data at scale without redundant decode work. Liked this video? Check out other Ray Summit breakout session recordings: https://www.youtube.com/playlist?list=PLZNaJyYZRPdo Subscribe to our YouTube channel to stay up-to-date on the future of AI: https://www.youtube.com/c/anyscale š Connect with us: LinkedIn: https://www.linkedin.com/company/joinanyscale/ X: https://x.com/anyscalecompute Website: https://www.anyscale.com/