Bae Jae-min

A South Korean software engineer and generative AI startup executive active in the industrial application of natural language processing.

Bae Jae-min

Overview

Bae Jae-min (배재민) is a South Korean software engineer and entrepreneur in the field of artificial intelligence (AI), known as a figure who has led the industrial application of natural language processing (NLP) and generative AI services. After majoring in computer science, he began his career as a search and recommendation systems developer, and later founded a technology startup developing conversational AI and large language model (LLM)-based services, drawing attention in South Korea's AI ecosystem. At the same time, "Bae Jae-min" is a relatively common name in South Korea, and there are namesakes with the same name in various fields, including academia, medicine, and culture and the arts.

Main Content

Upbringing and Education

Bae Jae-min is said to have been born in the early 1990s, and is known to have taken a strong interest in computers and programming from a young age. During his school years, he stood out by participating in programming competitions such as the Informatics Olympiad, and went on to enter the computer science department of a domestic university, where he studied natural language processing and machine learning intensively. It is said that in graduate school he conducted research on Korean morphological analysis and sentence embeddings.

Career

After graduating, he joined the search platform organization of a large domestic IT company, where he was responsible for query understanding and document ranking model development. He then passed through organizations handling recommendation systems and chatbot engines, accumulating experience in machine learning operations (MLOps) at the scale of live services. In the late 2010s, he founded a startup developing conversational AI solutions together with fellow developers, making enterprise customer service automation and document summarization services its flagship products.

Major Research and Development Achievements

  • Development of Korean-specific pretrained language models and research on lightweighting (distillation, quantization)
  • Construction of enterprise knowledge search systems based on retrieval-augmented generation (RAG)
  • Commercial service application of multilingual translation and summarization pipelines
  • Contributions to open-source Korean NLP libraries and release of datasets

These achievements have frequently been cited as reference cases in the process of domestic companies adopting generative AI in actual work.

Management Philosophy and Activities

He has stated in various lectures that reliability and safety must be secured alongside technical performance. In particular, he emphasized that reducing the side effects of generative AI—such as suppressing hallucination, protecting personal information, and copyright issues—is a precondition for industrial diffusion. He is also known to have participated in talent development activities, including mentoring in developer communities and universities and sponsoring education programs for young developers.

Namesakes

The name "Bae Jae-min" is used relatively frequently relative to the population, and there are figures with the same name in various fields, such as athletes, medical professionals, professors, and artists. Therefore, when citing documents or search results, it is recommended to confirm affiliation and field of activity together.

Latest Trends

Since 2024, the domestic and international AI industry has shifted its center of gravity from the advancement of large language models to "application and cost efficiency." In this trend, within the field to which Bae Jae-min belongs, on-site application of small language models (sLLMs), on-device AI, and the refinement of retrieval-augmented generation have emerged as key keywords. In 2024, as enterprise AI adoption began in earnest, capabilities for responding to security and regulation became important, and in 2025, along with discussions on institutional improvements such as the AI Framework Act, practical issues such as indicating the sources of training data and the obligation to label generated content have come to the fore. In addition, due to competition for securing GPUs and the problem of power costs, the value of model lightweighting and inference optimization technologies has grown further, and the need for models specialized in the Korean language and Korean cultural context is also being steadily raised. These changes are expanding the role of founders with backgrounds as natural language processing engineers from technology development to regulatory response and ecosystem building.

Related Topics

  • [[Artificial Intelligence]]
  • [[Generative AI]]
  • [[Natural Language Processing]]
  • [[Startups in South Korea]]
  • [[Large Language Models]]