This. R vs. Python is a very stupid debate. Especially for speed. For example pandas is slower than data.table. There are plenty of ways to have performant mode

This. R vs. Python is a very stupid debate. Especially for speed. For example pandas is slower than data.table. There are plenty of ways to have performant models in R. If you are worried about performance in deployment, run H2O, for example, there are plenty others. There is also flashr, if you want to write your own algorithms, which swaps R's basic operators and data management with C++. Also google is doing to R what they did to javqscript with v8. Expect GA next year.