What data scientists might not know:
Data that is collected and the ones presented for analysis has often a lot of preprocessing and transfer steps involved before it lands up in the data warehouse or analysis file. Most of the data scientists while learning ML / AI might have used already prepared data which eliminates the need for but in actual ML design in an industry often the data scientist has to prepare and modify data per the use case – they definitely need to know what was the data that was collected and how it ended up in a specific field (for example does Null gender mean that use did not want to share it or does it mean the data was unavailable or both – the data engineering team would have those answers)