Back to AI/ML & Deep Learning
About this interview
A technical interview on Neural Network Fundamentals, pitched at the medium level. A voice AI interviewer leads the conversation, adapts its questions to your answers, keeps you on topic, and afterward gives you honest, specific feedback on where you were strong and where to improve. Expect roughly 30 minutes.
What you'll be assessed on
Explain forward and backward propagation including the chain rule and vanishing/exploding gradients
Compare activation functions (Sigmoid, tanh, ReLU and variants) and their gradient properties
Describe gradient descent variants (SGD, Momentum, RMSprop, Adam) and their tradeoffs
Explain weight initialization strategies (Xavier, He) and normalization methods (BN, LN, GN)
Topics covered
Forward PropagationActivation FunctionsLoss FunctionsBackpropagationGradient Descent VariantsWeight InitializationNormalizationRegularizationOptimization LandscapeArchitecturePractical Training
A few sample questions
Just examples to set expectations - the real interview has many more and adapts to your responses.
“Can you walk me through what happens during a single forward pass through a fully connected neural network — from raw input to a final prediction?
“Why might SGD with momentum sometimes generalize better than Adam, even though Adam converges faster during training?
“What is the softmax function and when do you use it as an output activation? How does it differ from using a sigmoid for multi-class problems?