13.7 Coding and ML theory interviews
You clear the screening rounds.
Screening rounds are their own skill: Python data structures at the level actually asked, live Pandas and SQL drills, implementing k-means or logistic regression or attention from scratch, and crisp answers on bias-variance and metrics. It sits beside system design as the other interview gate. The misallocation to avoid is grinding hundreds of the hardest puzzle problems — ML screens rarely go there, while from-scratch implementations come up constantly and are practised far less.
Work through these
Python DSA at the level actually asked
Programming questions at the level actually asked in these interviews, which is usually less extreme than reputation suggests. Practising the right level saves months.
Pandas/SQL live-coding drills
Live data manipulation in a dataframe library and in a query language, under time pressure. This is where many otherwise strong candidates lose ground.
Implement from scratch: k-means, logistic regression, attention
Writing three standard algorithms from nothing, which tests understanding rather than recall. These three come up repeatedly.
Theory questions on bias-variance, regularization, metrics
Conceptual questions on the trade between simplicity and flexibility, on penalties, and on choosing measures. Being able to explain rather than define is what is being checked.
Sign in to keep your progress.
Free resources
We haven't checked most of these for screen reader use yet.
Links last checked 29 Aug 2026.
Stuck here?
Ask a mentor. A real person answers, and they can see exactly which topic you're on. Usually within a couple of working days.
Checking your session…
Topics shown in module order.