diff --git a/examples/mlx_demo.py b/examples/mlx_demo.py
index e69de29..304df39 100644
--- a/examples/mlx_demo.py
+++ b/examples/mlx_demo.py
@@ -0,0 +1,113 @@
+#!/usr/bin/env python3
+"""
+MLX Quick Demo
+==============
+
+A quick demonstration of MLX capabilities that work out of the box.
+"""
+
+import mlx.core as mx
+import numpy as np
+
+
+def main():
+    print("🍎 MLX Quick Demo - Apple Silicon Machine Learning")
+    print("=" * 60)
+    
+    print("\n1. 🎯 Basic Array Operations")
+    print("-" * 30)
+    a = mx.array([1, 2, 3, 4, 5])
+    b = mx.array([2, 3, 4, 5, 6])
+    result = a * b + 10
+    print(f"a = {a}")
+    print(f"b = {b}")
+    print(f"a * b + 10 = {result}")
+    
+    print("\n2. 🧮 Matrix Multiplication")
+    print("-" * 30)
+    A = mx.array([[1, 2], [3, 4]], dtype=mx.float32)
+    B = mx.array([[5, 6], [7, 8]], dtype=mx.float32)
+    C = A @ B
+    print(f"A =\n{A}")
+    print(f"B =\n{B}")
+    print(f"A @ B =\n{C}")
+    
+    print("\n3. 🔬 Scientific Computing")
+    print("-" * 30)
+    x = mx.linspace(0, 2*mx.pi, 10)
+    y = mx.sin(x)
+    print(f"x = linspace(0, 2π, 10)")
+    print(f"sin(x) = {y}")
+    
+    print("\n4. 📊 Statistics")
+    print("-" * 30)
+    data = mx.random.normal((1000,))
+    print(f"Random normal data (1000 samples):")
+    print(f"Mean: {mx.mean(data):.3f}")
+    print(f"Std:  {mx.std(data):.3f}")
+    print(f"Min:  {mx.min(data):.3f}")
+    print(f"Max:  {mx.max(data):.3f}")
+    
+    print("\n5. 🔄 NumPy Interoperability")
+    print("-" * 30)
+    np_array = np.array([1, 4, 9, 16, 25])
+    mlx_array = mx.array(np_array)
+    sqrt_mlx = mx.sqrt(mlx_array.astype(mx.float32))
+    back_to_numpy = np.array(sqrt_mlx)
+    
+    print(f"NumPy:     {np_array}")
+    print(f"MLX:       {mlx_array}")
+    print(f"sqrt(MLX): {sqrt_mlx}")
+    print(f"Back to NumPy: {back_to_numpy}")
+    
+    print("\n6. 🧠 Simple Neural Network Layer")
+    print("-" * 30)
+    # Input data (batch_size=3, features=4)
+    X = mx.random.normal((3, 4))
+    
+    # Weight matrix (features=4, outputs=2)
+    W = mx.random.normal((4, 2)) * 0.1
+    b = mx.zeros((2,))
+    
+    # Forward pass: linear layer
+    linear_out = X @ W + b
+    
+    # ReLU activation
+    relu_out = mx.maximum(linear_out, 0)
+    
+    print(f"Input shape: {X.shape}")
+    print(f"Weight shape: {W.shape}")
+    print(f"Output shape: {relu_out.shape}")
+    print(f"Sample output:\n{relu_out}")
+    
+    print("\n7. 🎨 Image-like Operations")
+    print("-" * 30)
+    # Create a synthetic "image"
+    height, width = 8, 8
+    image = mx.random.uniform(0, 1, (height, width))
+    
+    # Simple transformations
+    flipped = image[:, ::-1]  # Horizontal flip
+    cropped = image[2:6, 2:6]  # Center crop
+    
+    print(f"Original image shape: {image.shape}")
+    print(f"Flipped shape: {flipped.shape}")
+    print(f"Cropped shape: {cropped.shape}")
+    print(f"Image mean: {mx.mean(image):.3f}")
+    
+    print("\n8. 🚀 Performance Features")
+    print("-" * 30)
+    print("✅ Unified Memory: CPU and GPU share memory")
+    print("✅ Lazy Evaluation: Operations are fused for efficiency")
+    print("✅ Apple Silicon Optimized: Uses Metal Performance Shaders")
+    print("✅ Automatic Differentiation Ready")
+    print("✅ NumPy Compatible API")
+    
+    print("\n" + "=" * 60)
+    print("🎉 MLX Demo Complete!")
+    print("Ready to build ML applications on Apple Silicon!")
+    print("=" * 60)
+
+
+if __name__ == "__main__":
+    main()