Your question is Handle Vanishing Gradients. 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 is the vanishing gradients problem, and how would you handle it?
Explain the mechanism mathematically, show how you would detect it during training, and implement a small diagnostic experiment comparing an unstable deep network with a mitigated version. Discuss when initialization, activation choice, normalization, residual connections, learning-rate changes, or gated architectures are appropriate. Your answer should distinguish vanishing gradients from exploding gradients and optimization failure.