Datacurve increases $ 15 million to take over the AI ​​scale

As AI matures, the struggle for high quality data has become one of the most competitive sectors in industry, launching companies such as Mercor, Surge and, primarily, the AI ​​scale by Alexandr Wang. But now that Wang has gone to run AI at META, many financiers are looking at an opening – and are willing to fund companies with forcing new training strategies to collect training data.

The graduate of the combination y Datacurve It is such a company, focusing on high quality data on software development. On Thursday, the company announced a $ 15 million A -series round, led by Mark Goldberg in Chemistry with employees in Deepmind, Vercel, Anthropic and Openai. Series A comes after a $ 2.7 million seed round, which has been investing by former Coinbase CTO Balaji Srinivasan.

Datacurve uses a “Hunter Bounty” system to attract specialized software engineers to complete the most difficult data sets. The company pays for these contributions, distributing more than $ 1 million so far.

But co -founder Serena Ge (depicted above with co -founder Charley Lee) says that the biggest motivation is not economical. For high-value services such as software development, pay will always be much lower for data work than conventional employment-so the most important advantage of the company is a positive user experience.

“We treat this as a consumer product, not as a data marking feature,” GE said. “We spend a lot of time to think about: How can we optimize it so that the people we want to care about and get on our platform?”

This is especially important as the needs of after training data increases more complex. While previous models were trained in simple data sets, today’s AI products are based on complex RL environments, which must be manufactured through specific and strategies of data collection. As the environments grow more sophisticated, the data requirements are becoming the more intense for both the quantity and the quality-a factor that could provide high quality data collection companies such as datacurve an advantage.

As an early stage company, Datacurve focuses on software engineering, but GE says the model could apply equally easily on fields such as funding, marketing or even a drug.

“What we do now is that we are creating an infrastructure for collecting data after training that attracts and maintains extremely capable people in their own areas,” says GE.