Scale AI pushes into the synthetic data game with early access program

Scale AI pushes into the synthetic data game with early access program

by Pranali Mehta

San Francisco-based tech startup, Scale AI, is reportedly venturing into the synthetic data game, an emerging category in the AI field, as the company reached a $7.3 billion valuation through its latest funding round.

The firm announced an early access program for its product, Scale Synthetic, which allows machine learning (ML) engineers to enhance their existing real-world data sets.

For the building of this new division Scale also hired two new executives, Joel Kronander, who will be the head of synthetic data, and Vivek Raju Muppalla, who will be the director of synthetic services.

Synthetic data is fake data that is created with the help of machine learning algorithms, instead of using information collected from the real world, and is a powerful and handy tool for data generation, such as in medical imaging, where privacy is a major concern.

With synthetic data, developers will be able to add more complexity to their training models while removing biases that may be found in real-world data sets.

Scale had initially integrated software with real images, voice, text, and video data that was marked by people to offer autonomous vehicle enterprises the labelled information required to train ML (machine learning) models to develop and rollout autonomous vehicles. These vehicles included everything from  self-driving trucks, to robotaxis, and automated bots that are used in warehouses and for on-demand delivery.

Since then, the startup has transformed into a data management platform firm serving a wide range of customers, including government, e-commerce, autonomous vehicle, finance, and enterprise industries.

Alexandr Wang, founder and CEO of Scale AI, called the firm’s new product a hybrid approach to data, where they start with real data and build the product by growing and iterating it.

Wang explained that the company offered a powerful and unique offering to its customers by having real-world data as the base for synthetic data.

Wang also added that the U.S Department of Defense, Tractable AI, and Kodiak Robotics, have struck a deal with Scale for its synthetic data product.

Source credit:

Pranali Mehta

A chemical engineer by qualification, Pranali Mehta dutifully walked down the slated path and worked in a chemical firm for a year. Her passion for writing however, pushed her into experimenting with the same as a career. With over three years of experience in content writ Read more...