Large-scale AI has acquired a small startup which could help it expand its reach in Europe and accelerate the development of its new product.
Financial terms were not disclosed.
SiaSearch, from European venture capital studio Merantix, has developed a data management platform that acts as a search engine for petabyte-scale data captured by advanced driver assistance systems and automated driving. The startup, which already works with car manufacturers like Volkswagen and Porsche, is able to automatically index and structure raw sensor data collected by vehicle fleets.
This capability integrates seamlessly with existing Scale AI technology. Scale uses software and people to tag image, text, voice, and video data for companies that create machine learning algorithms. It was originally launched to provide autonomous vehicle companies with the labeled data needed to train machine learning models to develop and deploy axis robots, autonomous trucks and automated robots used in warehouses and delivery to the demand. The business has long since expanded beyond data labeling and is now more of a data management platform. It also serves other industries such as government, finance, e-commerce, and business and now works with companies like Airbnb, DoorDash, and Pinterest.
Berlin-based SiaSearch could be of particular benefit in building Nucleus, which Scale AI co-founder and CEO Alexandr Wang previously called “the first product of our future.” The plan is to integrate the team into the Nucleus effort, according to Wang.
Nucleus is an AI development platform that Wang describes as “Google Photos for Machine Learning Datasets.” The product provides customers with a way to organize, maintain, and manage large data sets, giving businesses a way to test their models and measure performance, among other tasks. SiaSearch enables Scale AI to accelerate its efforts and even expand functions to support the entire machine learning lifecycle, Wang said.
The goal is to integrate SiaSearch technology into Nucleus to provide a comprehensive data engine that any AI developer can use, even outside of automotive or audiovisual technology. This could prove to be extremely useful for any business – including robotics companies and automotive manufacturers – that not only needs to capture, tag and organize data, but also have additional tools to continually redefine new types of data. necessary to improve the algorithms used in its products. .
It sounds like what Tesla did, said Wang, who pointed out that the company is spearheading the concept of a data engine to help engineers improve the Autopilot advanced driver assistance system.
Wang said auto and robotics companies struggle to make the most of the vast amounts of data, especially as their fleets of vehicles, robots or other devices grow. Simply uploading all that data to the cloud would literally cost billions and billions of dollars, Wang said.
“Basically what every AI team is really looking for is how to energize our machine learning development and accelerate our dataset efforts, as much as Tesla was able to,” he said. declared. “We’re just going to give them the same superpowers as Tesla in terms of the ability to constantly overload their algorithms with the most relevant and interesting data from their mobile fleets.”