This video introduces this implementation utilized by Aupera and Xilinx on the VCK5000 card to achieve a full video processing and ML inference pipeline for AI recognition results with low latency and high throughput.