I haven’t used Ray much in my life, just a few times. Recently, when working on some LLM stuffy-stuff, I managed to find myself setting up a Ray cluster on Databricks for some distributed ML/AI work.
Lest you think I am some AI savant, tis not true, I still hold to the old axiom that 90% of all Machine Learning, including that fancy LLM stuff, is mostly the same ole’ Data Engineering.
So, what I want to do today is nothing fancy: introduce you to what Ray is, why, and where you would use it. Then, I will show you some code examples of how I used Ray on Databricks to fine-tune an LLM model … to help drive home the concepts of what Ray provides.