FastAPI is great building block but can't expect it to work for model serving out of box.
Load your pickle file and it works?
I know, pickle is insecure, etc
Most team don't really concern about the security side of things when using pickle for model deployment, more about performance and resource utilization. It works ok for very light weight models, but doesn't scale.
Actually it's very slow. Wait until you have 50 concurrent customers hitting it.
That's weird, it is supposed to be faster than alternatives. Do you have some benchmarks to support that critique?