Social media has transformed the way people form and express their opinions, making it an invaluable resource for sentiment analysis. This paper presents an in-depth opinion mining study, focusing on X (formally Twitter) data related to the Agnipath Recruitment Scheme in India, an initiative by the Indian government aimed at overhauling the recruitment process for the Indian Armed Forces. The tweets gauged the public’s reaction to the Agnipath Recruitment Scheme, understood the general sentiment (positive, negative, neutral), and identified key concerns or praises. By employing a hybrid model that blends Convolutional Neural Networks (CNN) and Bidirectional Long Short-Term Memory (BiLSTM) with eXtreme Gradient Boosting (XGBoost) classifiers, we were able to significantly increase the accuracy of sentiment classification. The findings of the study show that the feelings of the public are negative, and the maximum number of tweets on this topic is critical or dissatisfied with the public’s opinions on the Agnipath Recruitment Scheme. Our hybrid model, with its combination of sophisticated techniques, provides a strong method for deciphering and evaluating public mood on social media sites.