With the rapid growth of generative AI, an increasing number of companies are developing foundational vision models. These models demand billions of images for training and are significantly influenced by the quality of the input data and labels.
At Visual Layer, we assist Fortune 500 companies from various industries in curating their massive image and video datasets. In this presentation, I will summarize intriguing use cases we have encountered and share valuable lessons learned.
Danny Bickson, holds a Ph.D. in distributed systems from the Hebrew University and was a research scholar at Carnegie Mellon University. Danny co-founded GraphLab/Turi a Seattle based startup, acquired by Apple in 2016 becoming Apple’s deep learning infrastructure team. Danny’s recent startup, Visual Layer, is based on the popular open source project fastdup started a year ago by Danny and his collaborators Dr. Amir Alush and Prof. Carlos Guestrin from Stanford. Within a year fastdup reached 215,000 installations and was funded as a company by Madrona and Insight Partners. Visual Layer cloud product allows companies to manage their billion scale image and video datasets, driving insights and speeding up visual data ops.