Good place to start is datarobot themselves and one of their courses - see here https://www.datarobot.com/education/all-courses/. The Essentials course is actually focused very much on the basic principles rather than the technical details alone, and is great for understanding "why" rather than just how - which really helps I think!
The main thing though, is to just do a project from start to finish.
Outside of DataRobot, another excellent resource for beginners is Kaggle, a site that started as a place to host data science competitions and has now evolved into an community of users, example datasets, and free courses https://www.kaggle.com/.
I suggest you do the machine learning micro course for an overview then take on the titanic project. The titanic dataset is one that nearly everyone that has studied data science has looked at, don't know why, but it just lends itself very well to lots of missing data and a variety of types, perfect!
So Kaggle leans very much towards Python these days, which is actually very easy to get a grip of if you have some IT knowledge especially BI development. However, you mentioned you are already comfortable in R
I'd also highly recommend fast.ai from Jeremy Howard. Not only is he a leading researcher in the field, but he applies a top down learning which means you'll see (motivating) results / models immediately then fill in the theory as you progress.