AI Dataset Management startup Visual Layer has secured $7 million in seed funding led by Madrona and Insight Partners to accelerate the testing, training and fine-tuning of generative AI model management. Visual Layer’s technology, built around the open-source fastdub project, enables efficient identification and resolution of issues within visual datasets, ensuring optimal performance of AI models.
Visual Layer addresses the challenges faced by data scientists and machine learning engineers in managing and improving visual datasets used for AI model training by analyzing hundreds of millions of images, automatically identifying potential issues such as mislabeled, broken, or duplicate images. By streamlining the dataset management process, Visual Layer helps data professionals build more accurate and reliable AI models.
At the core of Visual Layer’s technology stack is the fastdub project, developed by co-founders Danny Bickson and Amir Alush, who have experience working with companies like Apple and Brodmann17. Fastdub leverages machine learning algorithms to identify and visualize clusters of potential issues within vast visual datasets. This open-source project provides a powerful tool for data scientists to ensure the quality and integrity of their image datasets.
The startup’s funding will enable services expansion and further develop its technology stack. The global AI training dataset market valued at $1.7 billion in 2022 and its expected to grow at a 22.1% CAGR from over the next seven year period to reach $8.7 billion.