##### Scenario A tech firm wants to assess a candidate’s grasp of ML and DL fundamentals for a new recommendation engine project. ##### Question Explain the difference between overfitting and underfitting and how to detect each. Name two advantages of deep learning over traditional machine-learning models and two disadvantages. Why are activation functions like ReLU preferred over sigmoid in deep networks? How do convolutional layers differ from fully connected layers in terms of parameter sharing and receptive field? ##### Hints Focus on general concepts—no math derivations required.