Re: [Question] use ML_PREDICT/ML_EVALUATE with python model provider in Flink

[Shengkai Fang]

I’m in favor of reusing the existing infrastructure as a starting
point, and refactoring it as the requirements evolve.

Best,

Shengkai