Datacurve Raises $15 Million To Take On Scaleai

As AI companies mature, nan conflict for high-quality information has go 1 of nan astir competitory areas successful nan industry, launching companies for illustration Mercor, Surge, and, astir prominently, Alexandr Wang’s ScaleAI. But now that Wang has moved to tally AI astatine Meta, galore funders spot an opening — and are consenting to money companies pinch compelling caller strategies for collecting training data.

The Y Combinator postgraduate Datacurve is 1 specified company, focusing connected high-quality information for package development. On Thursday, nan institution announced a $15 cardinal Series A round, led by Mark Goldberg astatine Chemistry pinch information from labor astatine DeepMind, Vercel, Anthropic, and OpenAI. The Series A comes aft a $2.7 cardinal seed round, which drew finance from erstwhile Coinbase CTO Balaji Srinivasan.

Datacurve uses a “bounty hunter” strategy to pull skilled package engineers to complete nan hardest-to-source datasets. The institution pays for those contributions, distributing complete $1 cardinal successful bounties truthful far.

But co-founder Serena Ge says nan biggest information isn’t financial. For high-value services for illustration package development, nan salary will ever beryllium acold little for information activity than accepted employment — truthful nan company’s astir important separator is simply a affirmative personification experience.

“We dainty this arsenic a user product, not a information labeling operation,” Ge said. “We walk a batch of clip reasoning about: How tin we optimize it truthful that nan group we want are willing and get onto our platform?”

That’s peculiarly important arsenic nan needs of post-training information turn much complex. While earlier models were trained connected elemental datasets, today’s AI products trust connected complex RL environments, which request to beryllium constructed done circumstantial and strategical information collection. As nan environments turn much sophisticated, nan information requirements go some much aggravated for some amount and value — a facet that could springiness high-quality information postulation companies for illustration Datacurve an edge.

As an early-stage company, Datacurve is focused connected package engineering, but Ge says nan exemplary could use conscionable arsenic easy to fields for illustration finance, marketing, aliases moreover medicine.

“What we’re doing correct now is we’re creating an infrastructure for station training information postulation that attracts and retains highly competent group successful their ain domains,” Ge says.