Optimizing a Hardware Network Stack to Realize an In-Network ML Inference Application | IEEE Conference Publication | IEEE Xplore