MediaZen Receives Pure Storage Customer Award... "Leading the Development of Diverse AI Products"

MediaZen announced on the 16th that it received the Customer Award at Pure Storage Accelerate 2024. Among the seven awardees worldwide, MediaZen was the only company selected from Korea.

The event took place on the 15th at Signiel Seoul in Songpa-gu, Seoul. Hosted by Pure Storage, a New York Stock Exchange-listed company, the award is given to customers who have achieved results using the Pure Storage platform. Approximately 150 people attended, including Lynn Lucas, Chief Marketing Officer (CMO) of Pure Storage, Jaesung Yoo, President of Pure Storage Korea, and Minkyu Song, CEO of MediaZen.

MediaZen is developing and supplying automotive, Artificial Intelligence Contact Center (AICC), AI education, and Large Language Model (LLM) technologies. In particular, its automotive voice recognition service is recognized in the industry for its high-quality competitiveness and technological capabilities.

CEO Song said, "To keep up with the rapidly advancing AI model research and development trends, we established an in-house Graphics Processing Unit (GPU) cluster to enhance our own R&D capabilities. Since a fast shared storage system that can exchange data between multiple GPUs is an essential element, we are using Pure Storage services."

He continued, "Previously, it was predicted that developing AI models based on large-scale voice data would take at least one year, but by utilizing Pure Storage, we completed training a voice recognition model with the targeted performance in about one month. Based on this, we were able to strengthen our position in the AICC sector."

The company plans to further intensify its research and development to create voice recognition models that meet all customer requirements. It is continuously researching the latest AI models through large amounts of data and GPU clusters, with the goal of achieving the desired voice recognition level for customers even with small amounts of data in the future.

CEO Song emphasized, "We are implementing technology that effectively removes only noise and extracts the voice of a specific speaker even in speech mixed with multiple languages or speakers," adding, "We will continue to strive to lead the development of various AI products."