Seoul-based startup Datumo has successfully raised $15.5 million in funding, aiming to strengthen its position in the competitive landscape of AI data solutions. David Kim, ex-AI specialist at Korea’s Agency for Defence Development, started up Datumo in 2018. He partnered with five other alumni from the Korea Advanced Institute of Science and Technology (KAIST) to get started on AI data labeling. This latest funding round nearly doubles the amount raised by the private company to about $28 million.
The story behind the genesis of Datumo’s creation originated from Kim’s experience and frustration with the long and cumbersome process of data labeling. So naturally, we started a company focused on data annotation services. In addition, we offer more sophisticated solutions for assessing AI models. “We first got our beginnings in data annotation, then becoming more involved in pretraining datasets and evaluation as this LLM ecosystem grew,” Kim said.
Datumo’s platform to evaluate the effectiveness of tests is a powerful, robust and agile evaluation platform called Datumo Eval. This unique new tool automatically generates test data. It tests responses against thousands of parameters to flag all unsafe, biased and factually inaccurate answers, removing the need for tedious manual scripting. These capabilities maximize the efficiency and accuracy of AI model assessments, something that is becoming more and more important in an AI-first economy.
The startup has been on a massive hiring spree and now employs over 150 employees in their Seoul headquarters. In March, Datumo established a presence in Silicon Valley, signaling its ambition to expand its reach beyond South Korea. The startup who’s anything but an underdog, currently serves over 300 clients. Included are the heavy hitters of South Korea’s tech economy including Samsung, LG Electronics, Hyundai, Naver and SK Telecom.
Datumo did around $6 million in revenue in 2024. In its first full year, she took the company to more than $1 million in revenue. The firm won some high-profile contracts right off the bat, helping to propel their trajectory to rapid success.
That funding round also brought in several high-profile investors, including Salesforce Ventures, KB Investment, ACVC Partners, and SBI Investment. These investments will accelerate Datumo’s mission to help transform drug development. They will provide our nation’s businesses with the standards, tools, and data they need to develop safe AI systems. The startup’s platform makes it easier to test, monitor and improve AI models. You don’t have to be a deep technical expert in order to engage with it.
Beyond those core offerings, Datumo has launched Korea’s first benchmark dataset for AI trust and safety. This latest dataset continues to position the company as a leader in promoting and advancing responsible AI practices.
Datumo stands out among its competitors through its licensed datasets, including one-of-a-kind data crawled from published books. This unique method significantly enhances the quality and reliability of its products. Consequently, it emerges as a desirable choice for companies wanting to improve their AI models.
Michael Hwang, involved with Datumo’s operations, elaborated on their service capabilities: “They wanted us to score their AI model outputs or compare them to other outputs.” It doesn’t help that clients are clamoring for data labeling services. They care deeply about having sophisticated evaluation tools to maximize the performance of their AI systems.