Machine learning engineering is hard, especially when developing products at high velocity (as is the case for us at Abnormal Security). Typical software engineering lifecycles often fail when developing ML systems.