Developed Tile operator support within the microsoft/onnxruntime repository, focusing on enhancing the WebNN Execution Provider’s capabilities for hardware-accelerated tensor operations. The work involved implementing the Tile operator using C++ and integrating it into the ONNX Runtime’s WebNN backend, thereby expanding the range of supported tensor operations and reducing reliance on CPU fallback for tile-heavy models. This addition improved execution paths for models leveraging the Tile operator, enabling better throughput on compatible devices. The integration was validated for correctness and prepared for broader testing, demonstrating proficiency in C++, ONNX Runtime internals, and the application of WebNN for efficient tensor computation.