This paper presents a study on gender identification and family relation extraction in the Hindu mythological text Srimad-Bhagavatam. The research is divided into two main components: gender identification and family relation identification. In the first part, we developed a model to determine the gender of characters using contextual cues, as many names alone are insufficient for gender classification. In the second part, we focused on extracting familial relationships between characters, identifying and categorizing relational phrases within sentences. We employed both traditional machine learning and advanced deep learning techniques, including Long Short-Term Memory (LSTM) networks and transformer-based models such as BERT, RoBERTa, and DistilBERT. Our results demonstrate that transformer models, especially BERT, significantly outperform traditional algorithms in both tasks. BERT showed superior precision, recall, and accuracy in gender identification and family relation extraction, underscoring the effectiveness of deep learning approaches in understanding and categorizing complex textual relationships in classical literature. This research provides valuable insights into the application of advanced NLP techniques for analyzing historical and mythological texts.