Interview
How Korea Deep Learning reached KRW 10 billion in cumulative revenue in five years without outside investment
Korea Deep LearningPublished

In an EBN interview published on April 22, 2026, Kim Ji-hyun, CEO of Korea Deep Learning, said the value of AI is proven not by the technology itself but by how much time it saves the people doing the work. Since its founding in 2019, Korea Deep Learning grew for its first five years on its own revenue alone, without outside investment, and passed KRW 10 billion in cumulative revenue five years after its founding.
Why Korea Deep Learning chose documents
Korea Deep Learning's reason for choosing documents was simple. No organization goes without digital documents, yet the AI that processes them was not good enough, the company judged. Its first business also came from documents. In 2019, its first year, it built a deep learning waybill recognition system for Korea Post and developed an ID card recognition module for Shinsegae I&C.
The gap between technology that only reads text and the technology the field demands was visible from then on. Where optical character recognition (OCR) technology in the past only recognized text, real industrial sites need technology that classifies specific values and understands context, in Korea Deep Learning's view.
Why it went five years without outside funding
For its first five years, Korea Deep Learning grew on its own earnings without outside funding. A KRW 100 million TIPS grant was the only money that came from outside. Korea Deep Learning judged that, ahead of the size of its capital, it first had to build technology that actually works well enough for customers to pay for it.
"If I couldn't create KRW 100 million in value with KRW 1 million of my own capital, I doubted that taking KRW 10 billion in outside funding would let me build a KRW 100 billion unicorn."
— Kim Ji-hyun, CEO, Korea Deep Learning
The result of building this way was the close of its Series A on January 8, 2026. Korea Development Bank (KDB) came in as the last investor with KRW 5 billion, closing a round of KRW 12 billion in total. It was the first large investment the company had received since its founding.
Waybill automation that shortened a courier's day
Korea Deep Learning defines AI as "a tool that gives people back their time." What matters in practice is not having OCR technology but whether it can turn documents into business data without error. Korea Deep Learning's yardstick is not whether a benchmark score is 99 or 95 but how much review time it saves the people doing the work.
The case Korea Deep Learning points to is the parcel delivery floor. It used AI to automate the work of couriers who had been working 10 to 11 hours a day typing in waybills by hand, and their job satisfaction rose. Korea Deep Learning counts it as its greatest reward to hear from the field that working hours fell and satisfaction rose after the solution was adopted.
This scene ties back to a problem Korea Deep Learning has worked on for a long time. In April 2022, Korea Deep Learning and Sunghwa Corporate Express were selected together for the AI Voucher Support Program of the National IT Industry Promotion Agency and developed a recognition system that reads handwritten waybills. The same year, its dual recognition system for shipping information, built with DEEP OCR, won the Korea Economic Daily President's Award at the 2022 Korea Impactech Awards. At the time, Korea Deep Learning explained that parcels pushed off sorting lines because their barcodes could not be read numbered at least 20,000 a day on average, and as many as 400,000.
When the field turns into conviction
Korea Deep Learning says its efforts turn into conviction when it sees its technology lighten the workload of people in the field and raise their satisfaction. That is why Korea Deep Learning sets its standard for technology in the field, not in scores.
The full interview was published in EBN.
About Korea Deep Learning
Korea Deep Learning (KDL), founded in 2019, is an enterprise AI agent company. Built on KDL Frontier, its own vision-language model (VLM) specialized for documents, KDL makes AI agents that take on and complete the document work of businesses and public institutions. Its main products are DEEP OCR, DEEP Parser, DEEP Agent and DEEP Ops.







