Developed and integrated K-EXAONE model support into the kvcache-ai/sglang repository, focusing on enabling causal language modeling tasks. The work involved updating model configuration files and introducing new inference-oriented model classes, allowing for streamlined experimentation and deployment of deep learning models. Leveraging Python and PyTorch, the developer aligned configuration and infrastructure to ensure the new model’s discoverability and readiness for production use. This integration established a foundation for expanded CLM capabilities within the codebase, supporting future model development and experimentation. The approach emphasized maintainability and cross-team collaboration, resulting in a robust addition to the machine learning workflow.