ML
ML Interview Roadmap
← Study plan·Math & Stats

Day 6 of 133

Sampling, bias, resampling — and pattern review

Random/stratified/cluster/importance sampling. Selection/survivorship bias.

DSA · NeetCode Sliding Window

  • Best Time TO Buy And Sell StockDSA · Sliding Window

    Interview questions to prep

    1. Walk through the one-pass min-so-far trick. Why is it correct?
    2. What if you can buy and sell unlimited times? What about with a cooldown or fee?
  • Longest Substring Without Repeating CharactersDSA · Sliding Window

    Interview questions to prep

    1. Why is this a sliding window? What's the invariant inside the window?
    2. What if the alphabet is huge (Unicode) vs tiny (ASCII)? Does that change your data structure?

Math · Sampling & survey design

  • Random, stratified, cluster, importance samplingStatisticsRead

    Interview questions to prep

    1. Compare simple random sampling vs stratified sampling — when does stratification reduce variance?
    2. What is importance sampling and where does it show up in ML?
  • Selection bias, survivorship bias, response biasStatisticsRead

    Interview questions to prep

    1. Give an example where selection bias silently broke a model deployment.
    2. What is survivorship bias and where does it hurt model evaluation?
  • Resampling for class imbalance & uncertaintyStatisticsimbalanced-learn

    Interview questions to prep

    1. What's the difference between bootstrapping for uncertainty vs resampling for imbalance?
    2. When does oversampling the minority class help and when does it leak?

References & further reading