← Meta Interview Insights

Meta·Software Engineer·Technical Phone Screen·Intermediate

Intermediate
Jun 2026

Summary

Quick ML theory question from a Meta interview, the kind that sounds straightforward but has a specific answer they're looking for.

Questions Asked (1)

Q1

In K-Nearest Neighbors, what happens to variance and bias when you set k=1?

Technical Trade-offsAlgorithms & Data Structures
Author's notes

k=1 means the model just memorizes every training point, so variance goes through the roof and bias drops to basically zero.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by defining bias and variance in the context of KNN, then explain how k=1 affects each. Conclude with the trade-off and practical implications.

Pro tip: Mention that k=1 leads to overfitting and high variance, but also note that it has low bias, which is a classic trade-off. Relate it to the bias-variance decomposition to show depth.

1. Define Bias and Variance

Briefly define bias as error from erroneous assumptions and variance as sensitivity to fluctuations in the training set.

2. Explain k=1 Behavior

With k=1, each prediction is based solely on the nearest neighbor, making the model highly sensitive to individual data points.

3. Analyze Variance

Variance is high because small changes in training data can drastically change predictions.

4. Analyze Bias

Bias is low because the model can fit complex patterns, but it may also fit noise.

5. Summarize Trade-off

Conclude that k=1 leads to overfitting: low bias but high variance, and discuss when this might be acceptable.

Key Points to Mention

  • Bias-variance trade-off
  • Overfitting
  • Sensitivity to noise
  • Model complexity
  • Effect of k on decision boundary
  • Practical considerations for choosing k

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.