Natural language processing in university tutoring management

University tutoring management faces complex challenges, such as high cultural diversity, the need for empathy, and the overload of inquiries in student tutoring sessions. To address these issues, an intelligent system was developed combining a lightweight backend server and an intuitive graphical interface to improve tutoring management. This system leverages Flask as the backend framework, enabling scalable web services, and PyQt5 to design an interactive graphical interface that facilitates monitoring and data management. Additionally, multithreaded programming ensures the simultaneous execution of the server and interface, improving user experience by preventing bottlenecks. The implemented method integrates advanced Natural Language Processing (NLP) algorithms, such as Naive Bayes and TF-IDF (Term Frequency-Inverse Document Frequency), to classify and extract relevant information, while Recurrent Neural Networks (RNNs) capture linguistic patterns in textual queries. These components work collaboratively through a backend API that communicates processed results to the interface in real-time. The development was carried out in Python, employing libraries such as NLTK, spaCy, and TensorFlow for language analysis and modeling. The system automated the tutoring process, reducing tutors' workload. With a 90% accuracy in intent classification and generated responses and an average response time of 1.2 seconds achieved through embeddings generated by Sentence-BERT, the system handled a higher volume of inquiries, increasing student satisfaction and optimizing tutors' time.