This document discusses machine learning techniques for diagnosing cardiac disease. It evaluates three datasets using different machine learning algorithms and proposes a custom convolutional neural network and extreme gradient boosting hybrid model that shows better accuracy. It also proposes a custom sequential dense neural network model with seven layers that achieves 92.3% accuracy on a modified Cleveland dataset for diagnosing cardiac disease. Previous related work applying machine learning methods like decision trees, K-nearest neighbors, and neural networks to cardiac disease diagnosis is also reviewed.