Implement a 3-Layer MLP for Linear Regression (Pure Numpy)

Question

### Problem: Implement a 3-Layer MLP for Linear Regression (Pure Python)

You are asked to implement a simple **multi-layer perceptron (MLP)** from scratch using only **Python and NumPy**.

### Specifications:
- The MLP has **3 layers**: input layer → hidden layer → output layer.
- Architecture:
  - Input dimension: **10**
  - Hidden layer 1: **64 neurons** (ReLU)
  - Hidden layer 2: **32 neurons** (ReLU)
  - Output layer: **1 neuron** (no activation)
- Loss function: **Mean Squared Error (MSE)**
- No use of any ML libraries (e.g., PyTorch, TensorFlow).
- You must implement both:
  - `forward(x)` for computing prediction
  - `backward(x, y)` for computing gradients

Accepted Answer

A detailed reference solution is available to Crack ML Interview members.