Comment by shrikar

The undifferentiated middle of ML engineering is getting squeezed. Your data rigor still wins, the experimentation loop gets a force multiplier, and here is a concrete 90-day plan to move.