Heaps, Top-K, and Binary Search

Three tools cover the streaming and search corner of an Adobe loop. A heap surfaces the k largest or most frequent items without sorting everything. Two heaps together track the median of a stream as numbers keep arriving. Binary search finds a value in a rotated array. It also pins down a monotonic answer when the input is never sorted, by halving the candidate range against a quick yes-or-no test. A Premium question and Kth Largest both come up a lot at Adobe. Rotated-array binary search is a recurring ask too. These three patterns earn their keep.

None of this is exotic. A heap is a priority queue you push to and pop the extreme from in log time. The two-heap median trick is one balancing rule applied on every insert. It keeps the two halves within one element of each other, so the middle is always one peek away. Binary search is the same halving loop you already know, with the comparison adjusted for a rotation or for a yes-or-no test on a candidate answer. Learn the four small skeletons below and this tier stops surprising you.