2025-12-052025-12-052025-06-23MELLO, Cláudio Klautau. Academick: assistente de estudos para o ambiente acadêmico. Orientador: Aldebaro Barreto da Rocha Klautau Júnior. 2025. 84f. Trabalho de Curso (Bacharelado em Engenharia Elétrica e Biomédica) – Faculdade de Engenharia Elétrica e Biomédica, Instituto de Tecnologia, Universidade Federal do Pará, Belém, 2025. Disponível em: https://bdm.ufpa.br/handle/prefix/8945. Acesso em:.https://bdm.ufpa.br/handle/prefix/8945The emergence of Large Language Models (LLMs) has created unprecedented opportunities to enhance educational processes, yet many higher education institutions have not fully leveraged these technologies. This thesis presents an adaptive educational chatbot system that integrates LLMs with Retrieval-Augmented Generation (RAG) to provide personalized learning assistance. The system’s key innovation lies in its intent-aware approach, utilizing a custom-trained classification model based on the Snowflake/snowflake-arctic-embed-l v2.0 architecture that achieved 97.6% accuracy on the test set. This classifier, trained on a bilingual dataset (Portuguese and English), enables the system to identify student intentions and dynamically adjust its response strategy. The system also features a novel PDF processing algorithm for efficient content extraction and indexing of educational materials. The chatbot can be rapidly deployed to support various educational scenarios, from textbook comprehension to academic article analysis, requiring only the source materials in PDF format and minimal setup configuration. By combining these compo nents, the chatbot provides context-aware responses, accurately cites source materials, and adapts its interaction style based on the identified student intent. This work demonstra tes how modern LLM technology can be effectively integrated into educational tools to enhance the learning experience while maintaining practical implementation considerations.Acesso AbertoGrandes Modelos de Linguagem (LLMs)Geração aumentada por recuperação (RAG)Tecnologia EducacionalProcessamento de linguagem naturalClassificação de intençõesChatbot educacionalAprendizagem adaptativaLarge language models (LLMs)Retrieval-augmented generation (RAG)Educational technologyNatural language processingIntent classificationEducational chatbotAdaptive learningCNPQ::ENGENHARIAS::ENGENHARIA ELETRICATrabalho de Curso - Graduação - MonografiaAttribution-NonCommercial-NoDerivs 3.0 Brazil