Crack Hulu’s 2020 Campus Coding Test: 4 Algorithm Challenges Explained

Hulu 2020 Campus Recruitment Programming Test Overview

The Hulu 2021 campus recruitment application phase has closed; HR is reviewing resumes and will notify candidates about the next steps. Whether you receive a direct interview invitation or need to take the online coding test, this guide offers the information you need.

Company Introduction

Q: What kind of company is Hulu? Hulu is a leading U.S. internet video streaming platform founded in 2006, now majority‑owned by Disney, with about 35 million paid subscribers and acclaimed original series.

Hulu values employee well‑being, offering competitive salaries, generous benefits, and a top‑ranked Beijing office.

Q: What does the Hulu Beijing office do? Established in 2007, it is Hulu’s second‑largest R&D center, responsible for video playback, recommendation systems, and targeted advertising, and it conducts research in artificial intelligence and machine learning.

Q: How will I learn about the online coding test schedule? After submitting your resume you will receive an email from HR by September 14 with the test details.

Online Coding Test Details

The test takes place on September 16 from 19:00 to 21:00 (2 hours) and is conducted remotely via a URL sent in the invitation email.

It consists of four questions: one easy warm‑up, two medium‑difficulty, and one hard problem. Each question has a detailed scoring rubric; partial credit is awarded for incomplete or sub‑optimal solutions.

Interview Tips

When you receive the problem : communicate actively with the interviewer.

When you encounter difficulty : think clearly, break the problem down, and stay calm.

When presenting a project : provide logical reasoning and avoid getting lost in minute details.

When coding : focus on code quality, style, and edge‑case handling rather than speed.

Problem 1 – Probability Josephus

Statement : n players (good = 1, bad = 0) stand in a circle. Starting from a randomly chosen player (weight w[i]), count clockwise from 1 to m; the player who says m is eliminated. Continue until one remains; the surviving player's team wins. Compute the probability that the good team wins, rounded to five decimal places.

Input : Line 1: n m Line 2: n integers a[i] (0 or 1) Line 3: n integers w[i]

Constraints : 0 < m < n ≤ 1000, 0 < ∑w[i] ≤ 10⁷

Output : a floating‑point number representing the good team’s winning probability.

Sample Input :

3 2
0 0 1
2 1 1

Sample Output : 0.50000 Analysis : The problem reduces to the classic Josephus problem. Compute the survivor index x for a start at 0 using the recurrence f(n,m) = (f(n‑1,m) + m) % n. The survivor for a start at i is (x + i) % n. Multiply the good‑team indicator array a by the weight array w after shifting by x, sum the products, and divide by the total weight. Both O(n²) simulation and O(n) recurrence are feasible for n ≤ 1000.

Problem 2 – Sum of Subarray Maximums

Statement : Given an array arr of video bitrates, define p[i][j] as the maximum value in the subarray arr[i…j]. Compute the sum of p[i][j] over all 0 ≤ i ≤ j < n, modulo 1 000 000 007.

Input : Line 1: n (1 ≤ n ≤ 10⁶) Line 2: n integers arr[i] (1 ≤ arr[i] ≤ 10⁶)

Output : the sum modulo 1 000 000 007.

Sample Input :

3
1 2 2

Sample Output : 11 Analysis : A naïve O(n²) scan computes each subarray maximum directly. Faster O(n log n) or O(n) solutions use a monotonic decreasing stack to find, for each element arr[i], the number of subarrays where it is the maximum: (i‑L+1)*(R‑i+1), where L and R are the nearest greater elements to the left and right. Multiplying this count by arr[i] and summing yields the answer in linear time.

Problem 3 – Minimum Walls to Remove in a Maze

Statement : A square matrix maze of size n × n contains 0 (path) and 1 (wall). Starting from the top‑left corner (0) you may move up, down, left, right on paths. Determine the minimum number of walls that must be turned into paths (1 → 0) so that a path exists to the bottom‑right corner (also 0).

Input : Line 1: n (2 ≤ n ≤ 5000) Next n lines: n integers (0 or 1) describing the maze.

Output : an integer – the minimal number of walls to remove.

Sample Input :

3
0 1 1
0 1 1
0 1 0

Sample Output : 1 Analysis : Perform a 0‑1 BFS (deque) where moving onto a 0 costs 0 and moving onto a 1 costs 1. The shortest‑path distance to the target gives the minimal number of walls to remove. This approach runs in O(n²) time and O(n²) memory.

Problem 4 – Maximizing Interest Value by Cutting a Match Schedule

Statement : A list of n matches (ids a₁…aₙ) may contain repeats. The list must be cut into k contiguous segments (preserving order). The interest value of a segment equals the number of distinct match ids within it. The total interest value is the sum over all segments. Compute the maximum possible total interest value.

Input : Line 1: n k (1 ≤ n ≤ 5000, 1 ≤ k ≤ 100, k ≤ n) Line 2: n integers aᵢ (1 ≤ aᵢ ≤ n)

Output : the maximum total interest value.

Sample Input :

3 2
2 1 1

Sample Output : 3 Analysis : Use dynamic programming where dp[i][j] is the best interest value for the first i matches cut into j segments. Let w[p][i] be the distinct count in segment p…i. The transition is dp[i][j] = max_{1≤p≤i} (dp[p‑1][j‑1] + w[p][i]). A naïve O(n²k) DP works for n ≤ 5000. To achieve O(n k log n), maintain a segment tree that supports adding +1 to a range when a new distinct id appears, allowing fast retrieval of the maximum dp[p‑1][j‑1] + w[p][i] for each i.