Lin Pyong enhanced the llvm/torch-mlir repository by implementing support for center_point_box=1 in the NonMaxSuppression operator, enabling seamless conversion between [x_center, y_center, width, height] and [x1, y1, x2, y2] box formats. Using C++ and MLIR, Lin developed and integrated end-to-end tests to validate the new behavior, ensuring robust compatibility for common object-detection workflows. This work reduced the need for manual format conversions upstream, streamlining the integration of object-detection models. The feature was delivered with a focus on reliability and maintainability, demonstrating depth in C++ development and a strong understanding of machine learning infrastructure requirements.