ML
ML Interview Roadmap
← Study plan·System Design

Day 107 of 133

Design TikTok / YouTube feed recommendations

Two-stage architecture; cold start; candidate-generation strategies.

DSA · NeetCode Stack

  • Min StackDSA · Stack

    Interview questions to prep

    1. How do you support getMin in O(1) — what's the auxiliary stack trick?
    2. Can you do it with a single stack of pairs, and what's the trade-off?

ML System Design · Recommendations

  • Design YouTube/TikTok feed recommendationsML System DesignGoogle

    Interview questions to prep

    1. Walk me through designing TikTok's For You feed end-to-end.
    2. How would you handle the cold-start problem for a new user?
    3. What candidate-generation strategies would you use, and how do you combine them?
    4. How would the architecture change for a Pinterest-style visual discovery or Follow Pins system?
    5. What feedback-loop or popularity-bias mitigation would you add for a visual discovery feed?
  • Two-stage candidate generation + rankingML System DesignGoogle

    Interview questions to prep

    1. Why do industrial recsys use a two-stage (recall + rank) architecture?
    2. How do you avoid training/serving skew between recall and ranking?

References & further reading