Moonsu Cha
Chief Technology Officer · Superb AI Inc.
Moonsu Cha is co-founder and Chief Technology Officer of Superb AI, the Seoul-based vision-AI company behind the Superb Platform. A former SK Telecom T-Brain research engineer and a co-author of the DiscoGAN paper presented at ICML 2017, he leads the company's technical direction as it moves from a one-stop MLOps and data-labeling platform toward zero-shot vision models and agentic vision AI for manufacturing, mobility, logistics and physical security.
Last reviewed: Sep 19, 2026
Moonsu Cha
Chief Technology Officer Superb AI Inc.
Last Reviewed: September 19, 2026
Executive Summary
Moonsu Cha is a co-founder of Superb AI and, since an appointment announced in February 2025, its Chief Technology Officer. Superb AI is a Seoul vision-AI company that began as a data-labeling business, became known internationally for that work, and then broadened into the Superb Platform — a single environment covering the whole machine-learning lifecycle from data construction, selection and annotation through model training, deployment and monitoring. Cha came to the company from SK Telecom’s T-Brain AI research group, where he was a co-author of DiscoGAN, one of the more widely cited pieces of Korean deep-learning research of its period. His argument for why the company stayed in front of the domestic industrial vision-AI market is that it sold a platform rather than a service: customers differed too much in demand, headcount and expertise for anything narrower to work.
Career Highlights
- Co-author of DiscoGAN, “Learning to Discover Cross-Domain Relations with Generative Adversarial Networks” (arXiv 1703.05192), presented at ICML 2017, as a research engineer at SK Telecom’s T-Brain group alongside Taeksoo Kim, Hyunsoo Kim, Jung Kwon Lee and Jiwon Kim.
- Co-founded Superb AI, which went through the Y Combinator Winter 2019 batch and is headquartered in Gangnam-gu, Seoul.
- Appointed Chief Technology Officer of Superb AI in an appointment announced in February 2025.
- Oversaw the company’s shift from data labeling to an all-in-one vision-AI adoption platform now holding more than 31 patents across four declared technology areas: no-code AI, data-centric AI, industrial AI and physical AI.
- Set the current technical direction toward zero-shot vision models and “agentic vision AI”, arguing that multimodal models, vertical AI and agents are a single connected problem rather than three separate ones.
- Superb AI has raised a $10.2 million Series C, bringing cumulative funding to $37.5 million, and had targeted a first-half-2026 listing on Korea’s technology track with Samsung Securities as lead underwriter; it remained unlisted as of September 2026.
Professional Journey
Cha’s technical reputation predates the company. At SK Telecom’s T-Brain research organisation he was second author on DiscoGAN, a 2017 paper showing how generative adversarial networks could learn relationships between two visual domains without paired examples. It was a well-timed piece of work — image-to-image translation was one of the hot problems of that moment — and it placed him alongside the group of SK Telecom researchers who went on to start Superb AI.
The company they built began with data. Labeling and data preparation were what Superb AI was known for, at home and abroad. But Cha’s account of the pivot is that customers kept exposing the limits of selling only that piece: the companies the team met in the field differed from one another in demand, in staffing and in expertise, and all of them struggled with the technology itself. The answer was the Superb Platform, which handles the full repeating MLOps cycle — data construction, selection, processing, management and analysis, then model training, deployment and operation — in one place. That completeness is what he credits for the speed at which vision AI reached Korean industry, and for the reception in Japan, where a Toyota representative noted that no platform bringing every MLOps function together existed locally.
His appointment as Chief Technology Officer was announced in February 2025, with the stated remit of lowering the barrier to AI adoption through customer-led product development. In the interviews he has given since, he has described the next stage as moving from a one-stop adoption platform to what he calls agentic vision AI: systems that act rather than wait, built on zero-shot models that do not need to be retrained for every new object or site. The commercial focus has narrowed at the same time, onto manufacturing, mobility, physical security and surveillance, with physical AI named as one of the company’s four core technology areas.
Superb AI has raised a $10.2 million Series C against $37.5 million cumulatively, and publicly aimed at a first-half-2026 technology-track listing in Korea with Samsung Securities as lead underwriter. That listing had not taken place by September 2026; the company’s investor-relations page at that date carried only private shareholder notices — a bonus issue in July and the withdrawal of a record date in August.
Education
No education record for Moonsu Cha was found in any source reviewed. A previously recorded pair of Ajou University degrees could not be corroborated and was removed; see the internal verification notes.
