You have worked on an NLP problem where you needed to turn raw text into a usable model input and ship something practical. The work likely involved choices around preprocessing, representation, model selection, and how to measure quality on messy real-world language.
Describe a project where you used NLP techniques. What challenges did you face?
How clearly you can explain an NLP problem and why it matteredWhether you can discuss tokenization and text preprocessing choicesHow you represented text, including embeddings or transformer featuresHow you evaluated the model, especially with F1 for imbalanced labels