Your question is NLP Preprocessing and Evaluation. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
How do you handle NLP data preprocessing, avoid overfitting, and choose evaluation metrics like F1 score?