During October 2025, Christi Babayeju developed a hardware-accelerated attention framework for the tathagatasrimani/codesign repository, focusing on optimizing deep learning computations for accelerator hardware. Christi implemented attention.cpp using C++ and integrated cache initialization and attention step processing to support efficient execution on FPGA platforms. The work included building an HLS benchmarking suite, enabling reproducible performance validation and early-stage optimization of attention mechanisms. By embedding these features into the Codesign pipeline, Christi established a foundation for future hardware acceleration and performance gains. The month’s efforts emphasized robust feature development and technical depth, with no major bugs reported during this initial integration phase.