Deploying Nvidia's Triton Inference Server with MinIO as the Model Store

The author writes really well. I would highly recommend his other writings as well.

The reason I like his writing is that he does some real world complex setups instead of demo hello world kind of examples.