Handshake acquires data-auditing startup Cleanlab

Handshake has acquired Cleanlab in an acqui-hire, which is defined as an acquisition primarily for talent. The transaction adds nine Cleanlab employees to Handshake’s research organization. This group includes Cleanlab’s co-founders, Curtis Northcutt, Jonas Mueller, and Anish Athalye, who earned their PhDs in computer science from MIT. The acquisition is designed to improve the quality of the data Handshake produces for artificial intelligence (AI) labs, specifically by integrating Cleanlab’s automated data-auditing expertise.

Cleanlab, which was founded in 2021, provides software for improving the quality of data produced by human labelers. Before the acquisition, Cleanlab raised a total of $30 million from investors, including Menlo Ventures, TQ Ventures, Bain Capital Ventures, and Databricks Ventures. At its peak, the startup had more than 30 employees. Handshake, which began in 2013 as a platform for hiring college graduates, launched its human data-labeling business about a year ago to serve foundational AI model companies. The platform now provides data for eight top AI labs, including OpenAI.

The financial details of Handshake’s growth and valuation include:

• Handshake was valued at $3.3 billion in 2022.
• The company was forecasted to end 2025 at $300 million in annualized revenue run rate (ARR).
• Handshake is reportedly on track to reach an ARR of “high hundreds of millions” this year.

Cleanlab’s researchers are experts in developing algorithms that flag incorrect data without requiring a second human reviewer. Sahil Bhaiwala, chief strategy and innovation officer at Handshake, explained that Handshake’s in-house research team thinks extensively about where their models are weak, what data they should be producing, and the quality of that data. Bhaiwala noted that the Cleanlab team has been focusing on this problem for years.

Northcutt, the Cleanlab CEO who is credited with pioneering its automated data-labeling auditing, noted that the startup received acquisition interest from other AI data-labeling companies. However, Cleanlab chose to sell to Handshake because competitors, including Mercor and Scale AI, frequently use Handshake’s platform to source human experts—such as doctors, lawyers, and scientists—for their data-labeling projects. “If you’re going to pick one, you should probably pick the source, not the middleman,” Northcutt said.

Why it matters

Handshake is doubling down on data quality by absorbing Cleanlab’s research team, aiming to improve the quality of data produced for AI labs.