During September 2025, Lizheng Zheng developed and integrated Qwen3-VL model support, including both standard and MoE variants, into the sgLang repository. This work expanded the platform’s capabilities to handle advanced vision-language tasks by updating the multimodal processing pipeline and introducing new configuration files. Lizheng applied deep learning and model integration expertise, leveraging Python to ensure seamless incorporation of the Qwen3-VL models within the existing framework. The engineering effort focused on enabling new multimodal AI functionalities without introducing regressions, demonstrating a solid understanding of both machine learning systems and practical integration challenges in production-level codebases. No bugs were reported.