Your experience highlights the balance between theory and practice in data science. One junior data scientist who took the Andrew Ng course gained a strong understanding of fundamentals, while the other, who focused on Kaggle, lacked insight into why certain modeling choices work. You stress that it’s not just about applying tools but understanding why they work. Without this, data scientists risk treating machine learning like a black box.
Kaggle is great for practical skills, but a solid theoretical foundation, like Ng’s course, is essential for structured decision-making. How have these differences impacted their ability to deliver results?
Some rookie data scientists manage to pass interviews without having a strong basis in the field, which is unexpected. Even juniors should learn the fundamentals because they will save time later. Although Kaggle teaches practical modeling (such as using sklearn), it is more difficult to provide the crucial explanation of why particular decisions are reached.
It’s also critical to explain results to stakeholders who are not technical. Beginners, particularly those switching from other professions, would benefit much from Andrew Ng’s course because he explains topics without making them too simple. It would be very beneficial to close the gap if juniors were encouraged to study fundamental subjects like regression, classification, and clustering.
This, not the course, is the difference, in my opinion.
Even Andrew Ng’s course doesn’t go very deep, but having a foundation in engineering and statistics gives you the fundamentals you need to undertake data science and machine learning.
Yes, particularly the section on statistics. With slightly different terminology, I would assume that this individual has at least performed logistic and linear regressions.