Parallel Domain develops a data-generation platform that creates virtual worlds and produces labeled sensor outputs (camera, radar and lidar) to train and test perception models for autonomy customers such as Google, Continental, Woven Planet and Toyota Research Institute. The platform offers a training mode that generates hundreds of thousands of parameterized examples and a testing API that lets customers place dynamic agents and integrate simulations with their own simulators. Parallel Domain automatically generates labels for each simulated agent to speed supervised learning and avoid manual labeling. The company plans to expand beyond automotive into drones and mobile computer vision and is doubling down on generative AI approaches to increase scenario and behavior diversity. It is hiring to support a growing customer base across North America, Europe and Asia. Recently the company closed a $30 million Series B to support these growth efforts. Parallel Domain provides a synthetic data generation platform composed of a suite of APIs and developer tools that generate synthetic sensor data on-demand. The platform is used through the full computer vision development cycle — from prototyping and model training to testing and post-deployment maintenance. Customers include leading AI companies, automakers, and delivery companies, who use the service to improve performance and reduce developer iteration time. Led by CEO and founder Kevin McNamara, the company focuses on reducing cost and time to deployment for accurate models. Parallel Domain intends to use its new funding to expand the footprint of its synthetic data generation platform. Parallel Domain develops simulation software that trains self-driving cars in virtual 3D worlds. Its platform uses real-world map data, procedural growth algorithms, and generative models to generate adjustable, programmable environments (lane count, terrain type, mountains, road curvature, etc.). The company plans to sell the platform to autonomous vehicle companies so they can generate virtual worlds in which to test and train their vehicles. Parallel Domain positions simulation as a way to reduce the danger of real-world training and to allow companies to focus engineering resources on building safe self-driving cars. Founder Kevin McNamara noted the urgency given test vehicles driving millions of miles per year. The company recently closed a $2.65 million seed round led by Costanoa Ventures, with participation from Ubiquity Ventures, Toyota AI Ventures and others.

Total raised

$44M

Funding rounds

3

Latest round

Series B

Latest activity

Nov 2022