A synthetic data API for AI developers. Data is the biggest pain point for AI teams. Data collection and labelling are expensive and tedious processes. We're building a synthetic data platform to help AI engineers create their own visual data quickly and affordably. Instead of manually collecting and labelling data, we want AI engineers to generate it. Users input what objects they want in their scene, as well as a number of environmental parameters. Our software assists users in generating photorealistic 3D scenes and extracting rich datasets that already come with industry-compatible labels. ""Give me a dataset of 10,000 segmented images of an urban street with 50% cars, 50% trucks in the winter at noon in Singapore, using the COCO annotation style"" We work with two main types of companies - (1) companies that want AI but have no data or expertise, and (2) companies with AI teams but insufficient data. We're working with SMRT for graffiti/anomaly detection on trains returning to depots, ST Engineering on license plate, model/make on cars in varying conditions, Aptiv on 3D virtual environments for autonomous vehicle testing, and Rolls Royce for AI turbine visual inspection.bifrost.ai
Middletown, DE, USA
Founded in 2019Inaccurate data? Flag it here.
Open jobs at Bifrost