Worked on the RosettaCommons/foundry repository to deliver end-to-end Intel XPU support for deep learning model workflows. Developed hardware-accelerated paths by introducing XPU-specific accelerator, precision, and strategy classes, enabling Foundry models to leverage Intel XPU for both training and inference. Enhanced distributed training utilities to automatically detect and utilize XPU hardware, and updated trainer configurations for RF3 and RFD3 models to streamline experimental workflows. Modified the MPNN inference engine for XPU compatibility, ensuring seamless integration. Provided comprehensive documentation updates to guide installation and usage. Utilized Python, PyTorch, and machine learning techniques to broaden deployment options and accelerate model experimentation.