Practical Builds

The Rippling onsite that people remember is a build. You get a short spec and a shared editor, and you ship a working service inside the hour while the interviewer runs it against cases and keeps adding to the spec. The three most-reported shapes are the same exercise in different clothes. One is an API you stand up in memory, one is a pile of records to group and roll up, and one is a grid you draw on. Each becomes a runnable problem below, so you practice the build itself.

One habit carries all three. Pick the plainest data structure that answers the query the spec asks for, get the first method working, and leave a seam where the next rule will land. A Q&A service is a couple of dictionaries and a counter. A group-and-aggregate is a bucket map and a sorted read. A canvas is a set of coordinates. None of that is clever, and that is the point: the interviewer is watching whether your first version stays correct as the spec grows.