Your question is Optimizing Deep Learning Models. 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).
What techniques would you use to optimize a deep learning model?
Discuss how you would diagnose whether the bottleneck is data quality, model architecture, optimization, regularization, or hardware utilization. Explain how you would run controlled experiments and evaluate improvements while avoiding overfitting and training-serving regressions.