This project focuses on predicting customer churn for a bank by classifying whether a customer will leave the bank (1) or stay (0). The main goal was to build a beginner-friendly deep learning workflow and understand how Keras and TensorFlow operate behind the scenes. I worked in google colab using kaggle inbuilt Keras interface, along with Pandas, NumPy, and Matplotlib for data preprocessing, analysis, and visualization. I developed two versions of the model: the first demonstrates the basic components of building a neural network, and the second includes improvements such as accuracy evaluation, validation split, threshold adjustments, and tuning epochs to enhance model performance. Through this project, I strengthened my understanding of neural network layers, training behavior, and overall deep learning workflow.