Vetiver Use Cases

Hi,

I have built a model in my work using tidymodels to identify fraud within our data.
It works so well I have been asked me to create it for another nine departments :slight_smile:

I only have my PC to build and train these models but recognize that maintaining 11 models on a weekly basis on top of my normal work might prove haphazard and time consuming. I am not an ML Ops expert by any stretch of the imagination but my setup is on windows. Modelling in general for me involves using the windows scheduler and comprises roughly of the following steps

Training every month

  • Pull the data from the data-warehouse/scrape from websites
  • Build the features - (tidyverse)
  • Train the model - (tidymodels)
  • Apply Sensitivity Analysis - (probably)
  • Save the model

Daily Run

  • Pull daily data, execute the model
  • Email results to the business (blastula + ggplot)

I have seen a recent post on vetiver which is MLOps in both R and Python

Is the expansion of my modelling task a good use case for vetiver?

I guess my requirements around this are basically to

  • Catch data drift
  • Catch a model failure
  • Make my life a bit easier for maintaining the modelling

I have no budget :slight_smile:

Thank you for your time

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.

If you have a query related to it or one of the replies, start a new topic and refer back with a link.