According to WHO estimates from 2016, diabetes has been named one of the chronic illnesses with the greatest rate of growth and is the seventh largest cause of mortality. For a clinically significant outcome, early identification of diabetes is always preferred due to the relatively long asymptomatic phase. Diabetes is a leading cause of stroke, heart attacks, kidney failure, blindness, and amputation of lower limbs. If it is discovered in its early stages, they could be prevented. The digital age has made an enormous quantity of clinical data available, which has allowed deep learning algorithms to provide accurate medical diagnosis and prognosis. To train the machine to predict it, we employ the diabetes dataset. The dataset was studied using the random forest, deep neural network, and logistic regression algorithms, together with a DNN that contained embedded data for the categorical features. With a F1 rating of 1.0 on test data, the DNN with insertions correctly predicts the majority of test set examples, or almost 100% of the cases.