Encord

London-based Encord provides a platform for curating, annotating and evaluating multimodal data that powers computer-vision and multimodal AI models.

It combines AI-assisted labeling with model-evaluation and human-in-the-loop tools, helping teams ship production-grade models faster. Used by hundreds of AI teams globally, it reported revenue growing roughly tenfold in a year.

It sits in the essential but underappreciated layer of data infrastructure beneath every serious vision model.