Hyundai Glovis pursued AI adoption to keep pace with rapid technological change in logistics and to automate and improve operations at its domestic KD centers and external C/C operations. To do this, it needed an integrated platform that connects to its existing GBP (Hadoop-based big data platform) and allows diverse AI development and experimentation in a short time, even in an air-gapped environment.

The key challenge was to make existing big data assets easy to use, enable fast and diverse AI development in an air-gapped environment, and respond flexibly to growing GPU demand.

With Cheetah, we built a Kubernetes-based AI development platform connected to GBP within one month, providing data access, version-specific development environments, and a foundation ready for future infrastructure expansion.

Hyundai Glovis can now develop AI more easily using existing GBP data and quickly set up purpose-built development environments, even in an air-gapped network. The Kubernetes-based architecture also provides a foundation for responding flexibly to future HPC infrastructure expansion and growing GPU demand.