Worked on the kvcache-ai/sglang repository to deliver the Intern-S1 Reasoning Parser and Tool-Calling Integration, enabling a new reasoning workflow and structured tool interactions. Focused on backend development using Python, the work included enhancing the grammar backend to robustly handle tokenization errors and implementing a detector for InternLM-formatted function calls, which streamlined tool invocation and improved automation reliability. Emphasized error handling and parser development to ensure consistent data flow and structured outputs. Collaborated with other contributors through co-authored commits, accelerating implementation and maintaining code quality. The feature addressed automation needs and increased the reliability of parsing and tool integration.