The integration of natural language processing (NLP) into educational information systems has proven highly advantageous for student interaction. Recent advancements in cognitive computing have introduced innovative ways to interact, facilitating quicker insights from existing data sources and enhancing the learning experience. This study explores the application of cognitive computing within blended learning environments. We propose a modular cognitive agent architecture designed for pedagogical question answering, enriched with social dialogue capabilities tailored to specific knowledge domains. Developed as a personal assistant, this system supports students in learning Data Science and Machine Learning techniques. Its implementation includes training machine learning models and refining natural language understanding algorithms within a user-friendly interface. The system's efficacy has been verified through experimental validation.

Natural Language Processing, Cognitive Computing, Blended Learning Environments, Pedagogical Question Answering, Social Dialogue, User-Friendly Interface, Machine Learning Models.