Thoughts: Domain knowledge isn't just the business domain.

The “domain knowledge” part of data science doesn’t necessarily need to be business domain knowledge. it could be the problem domain.

What do I mean?

When you’ve encountered a lot of problems, you start to see that at a fundamental level, these problems look like something you’ve seen before.

for example: churn prediction and machine failure at a fundamental level are all a part of the same problem domain, “make a prediction before something bad happens”.

Identifying different problem domains and having a set of principles to solve them is a great way to establish yourself as an expert outside of the business domain. This helps to generalise your skills across businesses and industries.