AI data-labeling company Handshake has acquired data quality startup Cleanlab, the two companies confirmed to TechCrunch. The acquisition brings Cleanlab’s research team and technology into Handshake as the company expands its services for foundational AI model developers.
Handshake, founded in 2013 as a hiring platform for college graduates, entered the AI data-labeling space roughly a year ago. Its newer business focuses on supplying high-quality human-labeled data to AI labs. Cleanlab, established in 2021, develops software designed to evaluate and improve the accuracy of data produced by human labelers.
The transaction is primarily structured as an acqui-hire. Nine members of Cleanlab’s team will join Handshake’s research organization, including co-founders Curtis Northcutt, Jonas Mueller, and Anish Athalye, all of whom hold PhDs in computer science from MIT. Financial terms were not disclosed.
Cleanlab had raised $30 million from investors such as Menlo Ventures, TQ Ventures, Bain Capital Ventures, and Databricks Ventures, and previously employed more than 30 people at its peak.
The startup is best known for its work on algorithms that automatically identify incorrect or low-quality labels without requiring a second human review. Handshake plans to integrate this research to improve the reliability of the datasets it produces for AI labs training large models.
“We have an in-house research team that thinks a lot about where our models are weak, what data should we be producing? How high quality is that data?” Sahil Bhaiwala, chief strategy and innovation officer at Handshake told TechCrunch. “The Cleanlab team has been focusing on this problem for years.”
According to Northcutt, Cleanlab attracted acquisition interest from several AI data-labeling companies before agreeing to the deal. He said the company ultimately chose Handshake because many data-labeling providers already rely on Handshake’s platform to source specialized human experts, including doctors, lawyers, and scientists, for their projects.
The acquisition positions Handshake to more tightly integrate human expertise with automated quality checks as competition intensifies around data quality in AI model training.
