Rippling — Software Engineer ✅ Passed

Level: Unknown Level

Round: Full Journey · Type: System Design · Difficulty: 7/10 · Duration: 60 min · Interviewer: Neutral

Topics: System Design, Data Structures, Algorithms, Hashmaps, Heaps, Concurrency

Location: San Francisco Bay Area

Interview date: 2026-01-06

Summary

Round 1: System Design

Question: Design a web API for an article voting system similar to Reddit or StackOverflow. The system should allow users to upvote or downvote articles, track vote flips, and rank articles. The get_most_recent_k_flips method must run in O(k) time.

---

Details

I was asked to design a system for article voting, similar to Reddit or StackOverflow, focusing on efficient data structures and time complexity.

The core requirement was to implement the following methods for an ArticleVotingSystem class:

` class ArticleVotingSystem: def add_article(self, article_name: str) -> int: """ Creates a new article. Returns a unique article_id. """ pass

def up_vote_article(self, article_id: int, user_id: int) -> None:
    """
    User votes UP (+1 score).
    - If they already voted UP: Do nothing.
    - If they voted DOWN before: Change it to UP (this is a "flip").
    """
    pass

def down_vote_article(self, article_id: int, user_id: int) -> None:
    """
    User votes DOWN (-1 score).
    - If they already voted DOWN: Do nothing.
    - If they voted UP before: Change it to DOWN (this is a "flip").
    """
    pass

def get_most_recent_k_flips(self, user_id: int, k: int) -> list[int]:
    """
    Get the last k articles where the user changed their mind ("flipped").
    A flip happens when you go Up -> Down OR Down -> Up.
    
    ⚠️ IMPORTANT: This must run in O(k) time, not O(total_flips).
    """
    pass

def get_top_k(self, k: int) -> list[int]:
    """
    Get the top k articles based on score (upvotes minus downvotes).
    The list should start with the highest score.
    """
    pass

`

Key constraints and rules included:

• Duplicate Votes: Ignore if a user votes twice in the same direction.
• Flip Definition: A flip is Up -> Down or Down -> Up.
• O(k) Time Complexity: get_most_recent_k_flips must run in O(k) time.
• Article IDs: Unique integers.

For example:

` system = ArticleVotingSystem()

Make articles

a1 = system.add_article("Python Tips") # ID: 1 a2 = system.add_article("Java Guide") # ID: 2 a3 = system.add_article("Go Patterns") # ID: 3 a4 = system.add_article("Rust Basics") # ID: 4

User 100 votes

system.up_vote_article(a1, user_id=100) # User 100: upvote 1 system.up_vote_article(a2, user_id=100) # User 100: upvote 2 system.down_vote_article(a1, user_id=100) # User 100: FLIP on 1 (up -> down) system.up_vote_article(a3, user_id=100) # User 100: upvote 3 system.down_vote_article(a2, user_id=100) # User 100: FLIP on 2 (up -> down) system.up_vote_article(a2, user_id=100) # User 100: FLIP on 2 (down -> up)

User 100's flip history: [article_1, article_2, article_2]

(Article 2 was flipped twice)

system.get_most_recent_k_flips(user_id=100, k=3)

Returns: [2, 2, 1] (Newest first)

system.get_most_recent_k_flips(user_id=100, k=2)

Returns: [2, 2]

Scores:

Article 1: -1 (upvoted then downvoted)

Article 2: +1 (up, then down, then up)

Article 3: +1 (upvoted once)

Article 4: 0 (no votes)

system.get_top_k(2) # Returns: [2, 3] or [3, 2] (both have score +1) `

I discussed using nested dictionaries for O(1) user vote lookups, a list to track vote flips for each user, and a heap for efficiently retrieving the top k articles. I also considered alternative data structures like deques and sorted lists to optimize get_most_recent_k_flips and get_top_k respectively. We also touched on handling vote deletions, tie-breaker scenarios, and concurrency considerations.