1) Enhancing communication and holistic optimization among application, algorithm, and platform development teams
2) Overcoming performance and power barriers of current platforms
3) Addressing scalability limits of model size
4) Ensuring technological reliability and performance of Korean language-based large-scale AI models