issue. There would have been a valuable equity piece to the compensation too. It was a startup with a very promising concept.

The data scientist in charge was insisting on using keys, and pseudo keys like timestamps from foreign systems as the primary keys in our systems. This was going to be for both their reporting warehouses and operational databases for customer facing websites.

I couldn't convince him it was going to lead to train wrecks. Even still, as the data engineer I knew it would be my head on a platter when things hit the fan. I gave them my recommendations and told them they needed to find someone who was more optimistic about their data scientist's unconventional modeling approach.