Deep learning techniques have gained a lot of traction in the field of NLP research. The aim of this paper is to predict the age and gender of an individual by inspecting their written text. We propose a supervised BERT-based classification technique in order to predict the age and gender of bloggers. The dataset used contains 681284 rows of data, with the information of the blogger's age, gender, and text of the blog written by them. We compare our algorithm to previous works in the same domain and achieve a better accuracy and F1 score. The accuracy reported for the prediction of age group was 84.2%, while the accuracy for the prediction of gender was 86.32%. This study relies on the raw capabilities of BERT to predict the classes of textual data efficiently. This paper shows promising capability in predicting the demographics of the author with high accuracy and can have wide applicability across multiple domains.