Neural Network Forward and Backward Pass Implementation

Problem Overview

Implement a simple multi-layer neural network entirely from scratch (no deep learning frameworks). You must write: (1) a forward pass computing pre-activations z = X·W + b, applying an activation function (e.g., sigmoid), and computing a scalar loss (e.g., MSE); (2) a backward pass applying the chain rule to derive…

• 4 candidate-reported follow-ups — the exact probes interviewers asked, with the trigger for each
• The rest of the problem statement — full requirements, constraints, and edge cases
• Approach and trade-offs — what passing candidates did, and the mistakes that sink people