This month, I set out to build on my previous work by building a ChatGPT pipeline from cloud data, using existing tables and developing a usable script.

Although I did not follow my exact weekly script or achieve the exact goals I set, my lofty aims allowed me to make significant progress in my understanding of what can be done. I learned how much processing needs to be done to ensure clean data and to generate all the functions needed for parameterizing generative AI.  Ultimately, I still did not directly connect to the database and pass live game information, but that remains a future goal.

As you can see below, there is significant data engineering and feature development needed in this process. The prompt is only a small part; in fact it’s just the tip of the iceberg.

On to September.