Worked on the vllm-project/vllm-ascend repository to deliver comprehensive deployment and validation documentation for Qwen3.5-27B and Qwen3.6-27B models. Focused on aligning documentation with vLLM v0.23.0, the work detailed prerequisites, installation, online service deployment for both single-node and multi-node environments, and performance tuning. Enhanced reproducibility by explicitly documenting post-processing parameters used in model validation, reducing onboarding time and clarifying evaluation workflows. Employed Markdown for technical writing and documentation, emphasizing model validation and software deployment best practices. The contributions improved production readiness and ensured cross-repository consistency, supporting scalable and reliable deployment in production environments.