3 ms·Millisecond Latency MLOps and the CERN for ML: An Interview with HuggingFace CTO2 points by thecellardoor 5y ago