Spark SQL

In the RDD API, you're exposed to extreme low level API's with Spark. So provided that you're comfortable with digging into source code and really understanding how a lot of that stuff works, you can do some pretty, pretty extremely low level things. You can directly access blast operators if you want to through Apache breeze. And you can implement your own version of a matrix transformation if you so choose. That's all open source. It's very hard to do that and to get it to be performing, but it's possible. Cool. I was thinking that we could potentially go through some of the key topics. And then as this is a machine learning podcast, start getting to more