The 60-second answer
Model log-odds as a linear function of features and map them to probabilities with sigmoid; multiclass commonly uses softmax. Train with negative log likelihood / cross-entropy and use regularization to control variance or induce sparsity.
Build the answer in this order
1
Give the core idea
Model log-odds as a linear function of features and map them to probabilities with sigmoid; multiclass commonly uses softmax.
2
Explain how it works
Train with negative log likelihood / cross-entropy and use regularization to control variance or induce sparsity.
3
Compare alternatives
Evaluate discrimination and calibration separately, then set thresholds from product error costs.
4
State failure modes + validation
Check feature scaling, collinearity, separation, leakage, and base-rate shift.
A useful interview mental model
This is the shape of a strong answer—not a script to memorize.
01Definition
02Mechanism
03Trade-offs
04Failure modes
05When to use
Senior-level signal
- Perfect separation can make unregularized maximum-likelihood coefficients diverge.
- In production, calibration and prevalence shift often matter more than small AUC differences.
What the interviewer is really testing
Mechanistic understanding, assumptions, trade-offs, and whether you can turn a definition into a model decision.
Likely follow-up questions
What assumption makes this approach work?
When would you choose the strongest alternative instead?
What production or data failure mode changes your answer?
Common weak-answer patterns
- Reciting a definition without mechanism or assumptions.
- Claiming one technique is always better without a data regime.
- Stopping before failure modes, validation, or deployment implications.