Most organizations say it is not ready to use ai generative in a safe and responsible way, According to the new MKKINSEY reportSee rankings-. One concern is a explanation – know how and why AI makes a particular decision. When 40% respondents see as a significant risk, only 17% active, each report.
Seoul basis Data Starting as a company’s data labeling company and now you want to help you make a business safer with a manner and data that activates tests, monitoring, and enhance technical expertise. The day of Monday starts $ 15.5 million, bringing a total performed by $ 28 million, from investors including Ventures Salesfor, and ACVC investment, among other investment, including.
David Kim, CEO Datumo and the former retaumo in the Korean body for Defense, frustrated using a new data that allows to make data available in their leisure and earn money. Startup persuaded the idea in the Introduction competition in Kaistive (Korean Institute of Science and Korean). Kim Co-feeded Datumo, who previously known as SecStar, along with Five Kais alumni in 2018.
Even before the app is built secure, sales pre-contract sales during the discovery of customers in competitions, usually of the led and startup business business.
In the first year, the starter beyond the $ 1 million income and safe some major contracts. Today, the beginning of calculating the main main company like Samsung, Samsung SDS, LG Electronics, LG CNS, Hyundai, Hyundai, Naver, and Seoul-based Telecomer base base. A few years ago, but the client began to ask the company beyond the easy data label. The first year of the age now has more than 300 clients in South Korea and returned about $ 6 million in 2024.
“They want us to print ai model output or compare to another output,” Michael Hwang, “Michael Hwang, co-fo-the estimates. “This is when we realized: We have done the Ai model evaluation – without knowing.” Datomo doubled in the area and release Korean first benchmarks focus on trust and security, hwn.
“We started on the data, then expanded it into pretraining datasets and evaluation as the ecosystem LLM resolved,” Kim tells her techcrroch.
TechCrunch Events
San Francisco
|
October 27, 2025
Meta $ 14.3 billion investment investment like investment in data-up-labeling company-labeling ai company highlight the importance of this market. Shortly after the affairs, AI Maker Models and meta competitors Opening stopped using Ai scale servicesSee rankings-. Deal meta also signifies that the AI training data competition.
Datumo shows several equations with companies like ai scale in the pretraining of the deciding density, and with Galileo and Arors ai The evaluation and monitoring of AI. However, distinguishes itself through the licensed datatet, especially the data that is released from the books published, which states the company offers rich with the armed humans but cannot be cleaned, according to COO Kim.
Not like friends, Datumo also offers a full-stack evaluation platform called EVAL dataAutomatically return data and automated evaluation automatically checks are not secure, the response is bias or wrong without the need of manual script, Kim add. The signature product is a code evaluation tool designed for non-developers as in policies, trust and safety, and a consolid team.
When asked at the interest of investors like the Salesforce Ventures, Kim explained that an opening before being a fire player with Andrew NG, founder of Deepearning.ai in South Korea. After that event, Kim showed sessions on LinkedIn, which made the attention of a salesforce. After some meetings and zoom calls, investors exceeded the soft commitment. The process of funding is all about eight months, Hwang said.
The new funding will be used to speed up the R & D effort, especially to develop automatic evaluations for the company AI, and in Japan, and also a bond in Seoul, which has 150 Silicon valley in March.