During July 2025, contributed to the pytorch/pytorch repository by implementing MTIA device support for foreach and fused kernels. This work involved expanding PyTorch’s device compatibility, allowing MTIA hardware to execute foreach and fused kernel operations, which can enhance performance and efficiency for MTIA-based machine learning workloads. The feature required a deep understanding of PyTorch’s backend development, particularly device subsystems and kernel dispatch mechanisms. Using Python as the primary language, the changes laid the foundation for hardware-accelerated pipelines on MTIA devices. No bugs were addressed during this period, with the focus remaining on feature development and integration within the PyTorch codebase.