Scaling R model to production

r
machine_learning
data_science

#1

What are all the ways to implement R model into production?


#2

Hi @Vikas,

good question. First which platform will you put in back end ? For example if you deal with near real time you stream kafka type or SQL then you need the API of some sort.
Your models will be RDS that have to be loaded with one common interface (input output). If you want to scale, deploy and destroy than put the RDS in docker, which is triggered by one scheduler and at last the monitoring.
Check this, it gives a good base.
It is not one easy task …
Best regards
Alain


#3

Thanks Alain,
I am dealing with realtime data
I am thinking of using Kafka, producers as webservices