
Original Coverage & Source Attribution: www.mk.co.kr
Nota announced on the 16th that it has signed an optimization large language model (LLM) and server supply contract with financial information technology (IT) company Dube and completed delivery.
Nota is a company specializing in artificial intelligence (AI) software that focuses on lightening AI models and optimizing reasoning and serving technologies.
Through this contract, Nota supports easy AI coding in the internal network established through its own facilities in the financial sector.
The financial sector has to use a closed network that restricts the external transmission of source code and internal business information during development, which limits the introduction of cloud-based AI coding tools.
The on-premise AI coding agent provided by Nota and Dube runs LLM on the customer’s internal server, and supports AI-based development tasks without calling external AI APIs, allowing companies to utilize AI coding tools while managing sensitive development information in their internal environment.
Through this contract, Nota will integrate and supply AI models from weight reduction to inference and serving environment optimization to server configuration in line with Dube’s AI coding agent.
In particular, in the same GPU environment, the number of requests for processing per minute increased by up to 95% compared to before the optimization was applied, and the performance improvement was confirmed by reducing the waiting time for the first response to less than half.
Based on this, it plans to support the introduction of AI agents in security-oriented corporate environments such as financial sectors where external data transmission is restricted.
Nota CEO Chae Myung-soo said, “We will support the efficient use of AI agents in environments with high security and infrastructure operational constraints such as the financial sector, and cooperate with various industrial partners to expand the business of supplying optimization software and AI execution environment.”
#