Worked on the pytorch/executorch repository to enhance the XNNPack backend by refactoring operation definitions and addressing alpha handling in mathematical operations. The refactor introduced wrapper macros for unary and binary operations, streamlining code and improving maintainability by removing redundant node definitions. Addressed correctness in add and sub operations by enforcing alpha constraints and ensuring proper fallback to portable operations when necessary. Developed comprehensive unit tests to verify these changes and prevent regressions. Utilized C++ and Python, focusing on backend development, performance optimization, and robust testing practices. The work improved code clarity, reliability, and long-term maintainability of the backend integration.