[AI Revolution](154) Complex AI Development at Once with Orchestration... Vesel AI

Vesel AI, a Company Specializing in AI Orchestration
Integrated Operations for Industry-Specific AI Models
Coordinating GPU, LLM, and Agents All at Once
"AI Transformation Will Change Every Job Within Three Years"

The core of an orchestra is the conductor. The conductor's role is to bring together dozens of different instruments to create a single melody. Likewise, artificial intelligence (AI) also needs a conductor. This is because the graphics processing unit (GPU) for AI computation, the large language model (LLM) responsible for natural language processing, and the AI agents that directly interact with users must work together as a team to deliver a service. The technology that integrates and coordinates this system as one is called 'AI orchestration.'

Jae Man Ahn, CEO of Vesel AI, is being interviewed by Asia Economy at the office in Seocho-gu, Seoul. Photo by Jo Yongjun

Vesel AI is a company that positions itself as an AI orchestration platform. Just as a conductor leads an orchestra, Vesel AI integrates multiple AI agents and provides them as a single service. Traditionally, each step from GPU adoption to LLM development and AI agent optimization had to be handled separately, but Vesel AI's integrated operations platform organically connects these processes. Jae Man Ahn, CEO of Vesel AI, founded the company in 2020, recognizing the importance of AI service infrastructure at the dawn of the AI boom.

As AI services proliferate these days, the importance of AI orchestration is growing even further. This is because it allows specialized AI services for different purposes to be bundled and used together. CEO Ahn said, "If you want to use AI to replace human work, it is difficult to do so with a single agent," and added, "Complex problems must be solved in the form of multi-agent systems." Agents specialized in each area, such as report writing, information retrieval, and data analysis, operate as if they are part of a single body.

Vesel AI also supports 'fine-tuning,' which allows companies to tailor AI models to their specific needs. Vesel AI's platform is based on open-source AI models such as Llama (Meta), DeepSeek, and Mistral. CEO Ahn explained, "There have been cases where banks fine-tuned models by training them with financial domain knowledge to create AI services for customer consultations, and in the United States, fine-tuned models specialized in specific academic fields such as science and life sciences are also emerging." If companies are concerned about data leakage during the AI model training process, they can also choose an 'on-premise' setup, where open-source AI models are built on their own servers.

CEO Ahn said, "Among Vesel AI's employees are engineers who previously managed infrastructure at global big tech and gaming companies," and added, "We have accumulated expertise in the AI field by training and fine-tuning LLMs over a long period." The company currently has about 30 employees and has also expanded to the United States, with an office in Silicon Valley.

CEO Ahn's goal is to expand AI orchestration to the extent that it can impact every industry in the global market. He predicted, "Within the next three years, every form of work we do will be transformed by AI," and added, "The starting point will be collaboration among multiple AI agents."

Bae Jae-man, CEO of Vesel AI, is being interviewed by Asia Economy at the office in Seocho-gu, Seoul. Photo by Jo Yongjun