You could probably keep 15 million rows in a csv file and still get great performance. Without even mentioning sqlite. Any reason to use such an overkill solution for that problem? Honest question, not passing judgement.
This was one of the better solutions they had. They also had Talend for ETL and Snowflake for “analysis”. You don’t want to know what Snowflake cost.
For the same roughly 15 millions rows…
It happened because he was the only semi technical person who was an FTE. Everyone else was from a consulting firm that the private equity owners “suggested” they use.