Personality traits provide insights into an individual’s behavior, thought patterns, and life state. User posts on social media are a good candidate for observing the individual’s behavior. A feature-enriched representation of textual content plays a crucial role in classification tasks. In the present work, a simple variant of max pooling, i.e., k-max pooling, is used to retain the essential sequential information for classification tasks. Additionally, BiLSTM is used to capture contextual information from both directions. The proposed model leverages the advantage of both approaches for extracting and retaining spatial, sequential, and contextual information for classification tasks. The average accuracy obtained with the base model is 51.78% and the proposed model achieves 71.38% accuracy on myPersonality benchmark datasets. Moreover, the comparison with the state-of-the-art models further demonstrates the effectiveness of the model proposed for personality traits prediction.