intelectron6/kyber-on-pynq ? reverse-engineered prompt

Build me a working CRYSTALS Kyber demo on a PYNQ Z2 board that splits the job between the FPGA and the ARM processor. I want the polynomial multiplication part accelerated in hardware using Vivado HLS, with the Zynq side handling the software, data movement, and control flow. Please set it up so I can run the Kyber flow end to end, from generating keys to encryption and decryption, and make it clear how the hardware block connects to the rest of the system.

If anything depends on setup details or current tool versions, look up the latest docs online and make reasonable choices. I’d also like simple build and run instructions so I can reproduce it on the board without having to figure everything out myself.