Examination assessments undertaken by educational institutions are pivotal since it is one of the fundamental steps to determining students' understanding and achievements for a distinct subject or course. Questions must be framed on the topics to meet the learning objectives and assess the student's capability in a particular subject. The generation of examination questions from extensive text material is challenging and complicated. For example, massive volumes of textbooks make it time-consuming for faculties to annotate good-quality questions, keeping them manually well balanced. Thus, teachers rely on the Bloom's taxonomy's cognitive domain, a popular framework to assess students' intellectual abilities. This study's motivation is to propose a pipeline that could provide new questions from a given text corpus that could be retrieved from a particular input. These generated questions could be incorporated into a question recommender while being automatically classified under the specific cognitive domain under the Bloom's taxonomy. Literature reviews showed that the work done over the Bloom's taxonomy domain had obtained results by implementing classical machine learning methods and few with deep neural networks. The proposed network architectures have shown remarkable results and state-of-the-art architectures compared to the literature. This research study concluded that the pipeline is effective and significant in generating questions, like manually drafting questions, categorizing them into the Bloom's taxonomy's domains, and providing explicit content-based question recommendations.