I don't have proper research experience, but my intuition tells me there are two key things for efficient research:

1. You want to reduce search space which is knowledge/experience.

2. You want traverse search space faster, i.e. reduce experiment iteration speed which is exactly engineering.