In lbartnik/chronicler: Binary History for R

library(knitr)
library(utilities)

knitr::opts_chunk$set(collapse = TRUE, comment = NA, prompt = FALSE, echo = TRUE)

Introduction

Machine Learning pipelines spotted in the wild seem to be designed more for production environments and less for the initial research. It shows in the data and processing models underlying the designs of those pipelines, and in the APIs exposed to its users. One prominent example of ML pipelines is the the ML Flow package which we will analyze in this document.

We will start with definitions of production environments and initial research and show why ML Flow is suited more for the former. Then we will describe a model of user experience suited better for the latter, implemented by the chronicler package.