I think "machine learning" in [this paper] applies fairly well to any type of scientific pipeline code:

> Using the framework of *technical debt*, we note that it is remarkably easy to incur massive ongoing maintenance costs at the system level when applying machine learning.

The authors argue that machine learning systems have the regular issues of a code, but also have other complexities that are not necessary addressed in the normal way of refactoring libraries, adding unit tests, etc.

[this paper]: https://research.google.com/pubs/pub43146.html "Machine Learning: The High-Interest Credit Card of Technical Debt"