Korea’s Datumo Raises $15.5M to Advance No-Code AI Evaluation for Trust and Safety

Seoul-based Datumo has raised $15.5 million in a new funding round, bringing its total capital raised to around $28 million. The investment was led by Salesforce Ventures, with participation from KB Investment, ACVC Partners, SBI Investment, and other backers. The company, founded in 2018 by five KAIST alumni, began as an AI data labeling startup but has evolved into a provider of tools for building safer, more responsible AI systems. David Kim, Datumo’s CEO and a former AI researcher at Korea’s Agency for Defense Development, was inspired to create the company after experiencing the inefficiencies of traditional data labeling. He developed a reward-based app that allows people to label data in their spare time and earn money. The concept gained traction early—during a KAIST startup competition, Datumo secured pre-contract sales worth tens of thousands of dollars from alumni-led startups and businesses, even before the app was fully built. In its first year, the company surpassed $1 million in revenue and landed key clients including Samsung, Samsung SDS, LG Electronics, LG CNS, Hyundai, Naver, and SK Telecom. Today, Datumo serves over 300 clients in South Korea and generated approximately $6 million in revenue in 2024. As client needs evolved, the company began receiving requests to evaluate AI model outputs, compare results, and assess performance. Michael Hwang, co-founder, noted that this shift revealed Datumo was already conducting AI model evaluation without realizing it. The company responded by launching Korea’s first benchmark dataset focused on AI trust and safety. Datuomo now offers a full-stack evaluation platform called Datumo Eval, which automatically generates test data and evaluations to detect unsafe, biased, or incorrect AI responses—without requiring manual coding. Its no-code interface is designed for non-technical teams such as policy, compliance, and trust and safety specialists. The company differentiates itself through its proprietary datasets, particularly those derived from published books, which provide rich, structured human reasoning. According to Kim, this data is valuable but notoriously difficult to clean and process. The recent $14.3 billion investment by Meta in Scale AI underscores the growing importance of training data and model evaluation in the AI industry. OpenAI’s decision to stop using Scale AI’s services shortly after the deal further highlights shifting dynamics in the space. While Datumo shares similarities with players like Scale AI, Galileo, and Arize AI, it stands out with its licensed datasets and automated, no-code evaluation tools. The new funding will support R&D in enterprise AI evaluation, expand global operations in South Korea, Japan, and the U.S., and strengthen its Silicon Valley presence, established in March. The fundraising process took about eight months, with Salesforce Ventures initially engaging after seeing a LinkedIn post from a fireside chat Kim hosted with Andrew Ng. The company, which employs 150 people in Seoul, continues to grow as demand increases for tools that help organizations deploy generative AI responsibly.