Given process or meeting intervals with inclusive endpoints, return the maximum number active at the same time. One version is a direct Meeting Rooms II-style prompt over `List > meetingTimings`; another describes distributed-service process logs.

Maximum Concurrent Processes / Meeting Rooms — IBM Interview Question

Input: a list of intervals, where each interval is [start, end].
Output: the maximum number of intervals active at any single time.
Intervals are closed: an interval ending at time end is still active at that exact time.
In the meeting-room version, the input is List<List<Integer>> meetingTimings; the return value is the minimum number of rooms needed.

Input: [[1, 3], [2, 4], [3, 6]]
Output: 3

At time 3, all three processes are active.

Input: [[2, 6], [1, 2], [3, 5]]
Output: 2

The sweep-line solution sorts starts and ends. Because intervals are inclusive, a start at time t must be processed before an end at time t when counting concurrency.
A min-heap of end times also works for the meeting-room framing; pop only ends that are strictly before the current inclusive start.
Difference-array / prefix-sum solutions are viable only when the time domain is small or coordinate-compressed. With times up to large values, sorted events or a heap are safer.
This is the same algorithmic family as Meeting Rooms II; the IBM variant mainly changes the input container, story, and endpoint convention.

Implement both versions: sorted event pairs with tie-breaking, and min-heap over end times. Add a test where one interval starts exactly when another ends.
Drill the complexity explanation: sorting dominates at O(n log n), heap space is O(n), and coordinate compression changes the prefix-sum version from time-domain-sized memory to O(n) events.