Korean startup Hyperlearning scales AI assessment to 6,000 students
Hyperlearning's Hyper-Rater has graded writing for 6,000 students at Konkuk University; the startup now targets the U.S. with a Berkeley partnership and AI exam tools.

Why it matters
- Hyper-Rater has evaluated approximately 6,000 students at Konkuk University's Glocal Campus over three years.
- Hyperlearning holds 11 academic papers and four patents supporting its assessment technology.
- In 2026 the company opened a branch in Johor Bahru, Malaysia, and partnered with UC Berkeley's BEAR Institute.
- The Korean Essay Test launched in Hanoi, Vietnam, in late 2025 for nonnative Korean writers.
- Products include Hyper-Rater, HyperQ for exam question generation, and the Dabanji grading platform.
Hyperlearning's AI grading system Hyper-Rater has evaluated approximately 6,000 students at Konkuk University's Glocal Campus over three years, and the Korean education technology startup is now preparing to enter the U.S. market. The company, which specializes in AI-powered assessment, has built a product line covering automated grading, examination question generation and personalized feedback, and it has expanded across Southeast Asia with a branch in Malaysia and a partnership in Singapore during 2026.
The stakes are straightforward for schools and universities. Teachers spend increasing amounts of time on grading and administrative work, which leaves less time for students, and the market for specialized educational assessment technology has relatively few providers. Hyperlearning's co-founder and CEO Lim Chae-won argues that AI should not merely accelerate existing assessment methods but should change how educators evaluate students in the first place.
"In the age of AI, the way we assess students and the very paradigm of educational assessment should be changed," Lim said in an interview with The Korea Times.
What does Hyperlearning actually sell?
The company's portfolio consists of three core products plus one certification examination:
- Hyper-Rater automatically evaluates written responses and provides personalized feedback to each student.
- HyperQ generates questions for high-stakes examinations, including Korea's College Scholastic Ability Test.
- Dabanji is an online platform for automated grading and test administration.
- Korean Essay Test (KET) is an online certification examination that evaluates the Korean writing proficiency of nonnative speakers.
The company says its technology rests on 11 academic papers and four patents. Lim frames the technical foundation as the difference between general-purpose AI and assessment work that demands fairness and consistency across large numbers of students. Generating high-quality examination questions and evaluating student responses fairly requires a deep understanding of education and assessment methodologies, he said, expertise that goes beyond what off-the-shelf language models deliver.
The leadership team adds another layer of credibility, according to Lim. He worked with major Korean education companies, including Etoos and Hiconsy, before founding Hyperlearning, and the team combines technical capabilities with practical knowledge of how schools operate.
Why did the founder build an assessment company?
Lim's motivation came from direct classroom experience. Before founding the company, he worked as a private school teacher and saw how much time educators spend on administrative responsibilities and assessment. Teachers evaluate students through written assignments, oral examinations and portfolio-based assessments, but when they carry responsibility for multiple classes, providing detailed feedback and managing grading workloads becomes extremely difficult.
Hyperlearning was established to address that gap through AI-powered assessment. The company's products aim to take over the repetitive parts of evaluation — scoring written responses, drafting feedback, generating test questions and administering examinations — so that teachers can redirect their hours toward students themselves.
Where is the technology in use today?
Hyper-Rater has run for three years in freshman writing diagnostic assessments at Konkuk University's Glocal Campus, covering approximately 6,000 students. Lim said the system has reduced the time required for assessment while providing consistent evaluations and allowing the university to monitor students' progress over time.
In the private education sector, Hyperlearning has conducted a proof-of-concept project with institutions in Daechi-dong, Seoul's best-known cram school district, to introduce its essay assessment technology.
International deployment began with KET. The company introduced the Korean writing test in Hanoi, Vietnam, in late 2025, allowing local Korean-language learners to assess their writing abilities and receive personalized results and feedback. Hyperlearning is now preparing to expand into other Southeast Asian markets, including Indonesia, where Lim said demand for Korean-language education is growing.
What happened in 2026?
The year split into a Southeast Asian expansion phase and a U.S. preparation phase.
In the first half of 2026, Hyperlearning expanded partnerships related to Korean-language assessment in Southeast Asia, including with the National University of Singapore. It also established a branch in Johor Bahru, Malaysia, to jointly operate an education center.
In the second half, the company joined the Global Market Expansion Program, supported by the Korea Institute of Startup and Entrepreneurship Development, to prepare for entry into the U.S. market. Hyperlearning introduced its business to investors in Silicon Valley and participated in EdTech Week 2026 in New York. Lim said those meetings surfaced growing interest in educational assessment technology and opportunities in a market with relatively few specialized providers.
The most consequential move came in academia. Hyperlearning established a strategic partnership with the BEAR Institute at the University of California, Berkeley's Graduate School of Education to explore U.S. market localization and collaboration in subject-specific, oral and SAT assessment.
What do U.S. universities want from AI assessment?
Lim's meetings with university provosts shaped his view of the American market. He found that provosts were generally open to adopting AI but wanted clear evidence that solutions were reliable and appropriate for educational use before implementing them institution-wide.
That finding points to the central challenge for every AI vendor in education: proof. Lim said educational assessment systems must be developed and validated through rigorous frameworks to establish trust and reliability, a requirement he ties to the growing importance of responsible AI.
He identified two structural shifts in how U.S. institutions think about assessment. First, as students increasingly use AI to complete assignments, educators are placing greater emphasis on problem-solving, critical thinking and creativity rather than simply evaluating final answers. Second, interest in oral assessment is growing because universities want more accurate ways to understand students' actual abilities beyond traditional written examinations.
Both trends favor companies that assess process and ability rather than final output alone, and both complicate the business of any vendor whose product grades only finished text.
What is the long-term plan?
Hyperlearning's stated goal is to build comprehensive assessment infrastructure covering the entire process, from question generation and grading to learning analytics. Lim said the company believes assessment can drive educational transformation in the age of AI, positioning evaluation — not content delivery — as the layer where AI changes education most.
The immediate test of that thesis comes in the United States. Hyperlearning enters a market where provosts demand validation evidence, where students already use AI on assignments and where the Berkeley partnership gives it a research foothold at a graduate school of education. Whether the company can convert 6,000 assessed students in Korea and a growing Southeast Asian footprint into institutional contracts in America will determine whether assessment infrastructure becomes a category or stays a niche.
Source: feed.koreatimes.co.kr
More from Rebecca Stone
Show full bio
Correspondent covering companies, deals and strategy in artificial intelligence and machine learning at AI In Context.
728 articles
Related articles
- A Former Professor Estimates 90% of His Online Students Used AI for Coursework
- Pearson to Acquire AI Skills Platform Workera in H2 2026
- Korea adds 8,400 staff for AI-cheating safeguards at CSAT
- AI Glasses Solved Four CSAT Math Questions in Ten Seconds
- Google DeepMind Launches ATL Saathi, a Gemini Assistant for Indian Teachers