Jang Hyun (Vincent) Cho announced on X on August 26 that he had joined Thinking Machines Lab the previous week. His new title was not disclosed.
Cho spent nearly two years at Meta, where he drove work on multimodal pretraining for the Muse and Spark model family and contributed to another project in the Segment Anything family. His broader research focuses on systems that understand images, language and 3D scenes.