把自信拆成准备、表达和复盘

Two timed Python problems; explain trade-offs clearly

Explain leakage, calibration, and confidence intervals

Design monitoring for model and data drift

Build an evaluation set, failure taxonomy, and tracing

Practice walk-forward validation and cost-aware evaluation

Use answer-first structure with one recommendation

Prepare six stories with measurable outcomes

Use concise answers, a calm pace, and deliberate pauses