ansys·Software Engineer·Technical Phone Screen
- How do you detect overfitting versus underfitting, and what do you do to fix each?
- What is the difference between regularization, standardization, and normalization, and when would you use each?
- Compare Random Forest and LightGBM. When would you choose one over the other?
- Explain the bias-variance trade-off and how model capacity, regularization, and dataset size each affect it.
- Walk me through cross-validation, train/val/test leakage, feature engineering, and evaluation metrics.
“I started with the training vs validation loss gap explanation which felt solid, but then I fumbled a bit listing fixes.”