Feature engineering is the main bottleneck (after data cleaning) in any machine learning pipeline. The sooner you’re done with it the sooner you can do other, cool stuff, which is why there are now so many solutions that focus on around feature engineering, whether it’s explicitly (feature stores) or implicitly (neural networks.)
It is impossible to stay onboarded without good async practices such as documentation. Documentation is not a luxury, it is a critical path for remote engineering work.