Explanation
- You need to remember `subarray(i, j) = prefix(i) - prefix(j)`, where `i >= j`
- relation reduced to `prefix(j) <= prefix(i) - k`
- relation reduced to `prefix(j) + k <= prefix(i)`
- at each jth value apply upper bound to count the values
- **corner case**, count each number having `value >= k` individually.