Arrays, Strings & Intervals

Snowflake's loud coding signatures are topological sort and build-a-data-structure. This lesson is the rest of the room: the sliding-window, two-pointer, hashing, heap, interval, and stack problems that fill out a phone screen or a second onsite slot. They are not the headline, but they show up, and a few of them show up a lot. A Premium question lands in Snowflake reports about as often as any string problem in the set. Merge k Sorted Lists, a Premium question, Trapping Rain Water, Top K Frequent, a Premium question: all recurring, all in this band.

These read like standard mediums and the occasional hard. The catch is what gets stapled on. Snowflake is a data warehouse, so a problem that looks like "find the shortest window" arrives as "aggregate this log without scanning it twice." The grade rides on whether your structure holds at a billion rows; reciting the algorithm is not enough. Same patterns everyone drills. A scale clause you have to hear.

The Snowflake angle

Restate the inputs, the scale, and the edge cases before you type. Candidates fail this loop by skipping it. Snowflake interviewers want the requirements pinned first, then correct code, then a defense of the time and space. A medium-looking prompt usually has a data-scale twist folded in, and diving straight at a solution misses it. Ask how big the input gets, whether it streams or sits in memory, whether values can repeat or go negative. Then pick the structure that answers the scale, not just the example; the first two minutes are graded.