Contributed core operator enhancements to the FlagOpen/FlagGems repository, focusing on numerical reliability and image processing capabilities. Developed and integrated an asinh operator with robust edge-case handling, ensuring numerical stability across FP16, BF16, and FP32 precisions and aligning behavior with PyTorch semantics. Added a pixel_shuffle operator to support advanced image manipulation workflows, organizing dedicated per-operator test files to improve maintainability and reliability. Emphasized comprehensive cross-precision validation and test coverage, reducing NaN and infinity edge-case failures. Collaborated with co-authors to deliver production-ready features using Python, PyTorch, and numerical analysis, supporting long-term operator development and downstream analytics workloads.