In the interviewer’s wording, not a cleaned-up version. Tagged by round, by level and by the track that covers it.
The reasoning, not an answer key. What you would say out loud, in what order, and where to stop.
The two or three questions that came next, which is where most candidates actually lose the round.
Depth, calibration, communication or judgment — knowing which one changes how you answer.
A question bank is only worth something if it reflects what is really being asked. Three rules keep it that way.
Entries come from members who sat the loop, written up afterwards. We do not scrape forums, and we do not pad the count with questions nobody has been asked.
Take-home contents, live coding platforms and anything a candidate agreed not to share stay out. Questions are described in the form they are asked, never lifted from confidential material.
Companies change their loops. An entry that stops being reported is marked historical rather than left to look current.
Graded on whether you reach for the cost asymmetry unprompted. The follow-up is always where you would put the threshold, and why.
Graded on whether the shapes are right and whether you narrate while writing. Partial code with correct reasoning beats silent complete code.
Graded on structure. Candidates who spend fifteen minutes on requirements never reach the tradeoffs, and the tradeoffs are the round.
Graded on whether training/serving skew is on your list, and how fast. The follow-up asks how you would have caught it earlier.
Graded on specificity and on whether you own the decision. A story with numbers in it survives three follow-ups; a tidy one does not.
The free track that teaches the five rounds. The question bank is the drill material for it.
Open the trackPro is $9,900, paid once, for 5 years of access. The question bank is part of it.
See the pricingSat a loop recently? Send us what you were asked and it goes into the bank.
Send it in
