work On Device LLM Serving Engine Optimizing and supporting language models for execution on NPU backend. Small Langauge Model Training small language models for on-device deployment. Korean Reasoning Model Training an o1-like large language model capable of reasoning and responding in Korean. Vocabulary Expansion Expanding vocabulary of large language model to include Korean language capabilities. Knowledge Distillation Transferring the capabilities of a large language model to a smaller one. Domain Adaptation Training a large language model capable of understanding domain-specific task.
