Nogueira, Marcelo BorgesNeves, Christopher Matheus Florencio Ramos2024-08-302024-08-302024-08-08NEVES, Christopher Matheus Florencio Ramos. CampusHubAI: Uma integração inovadora do aluno à universidade com GenAI.2024. 51 f. Trabalho de Conclusão de Curso (Graduação em Ciências e Tecnologia com Ênfase em Computação Aplicada), Escola de Ciências e Tecnologia, Universidade Federal do Rio Grande do Norte, Natal, 2024.https://repositorio.ufrn.br/handle/123456789/59957This work presents the concept of developing a platform called CampusHubAI, which aims to utilize Generative Artificial Intelligence (GenAI) to establish a natural language connection between members of the academic body and the institution's document database. This solution will enable users to address any inquiries regarding regulations, resolutions, and other data and documents from their respective educational institutions through easy machine interaction, via text or audio. Furthermore, the system will provide means of consultation, through conversations and web pages, for information on academic affiliations, enabling the analysis of grade history, access to educational materials, and other resources. This approach offers university members greater ease and flexibility in understanding documents and data, while also reducing the workload of human attendants. The Hub also has the potential to positively contribute by allowing introverted users to resolve their issues with minimal human interaction. The platform will be developed using JavaScript's Bun Runtime, utilizing the Elysia library for internal API, RabbitMQ for message and load control, the LangChain framework to accelerate the development of tools with GenAI, and Google Vertex AI for improved scalability and flexibility among the main models in the market, such as Claude 3.5 Sonnet, Llama 3.1, and Gemini 1.5 Pro. The Hub will employ the concept of Retrieval-Augmented Generation (RAG), combining the power of language models with a vector database using Apache Cassandra. Additionally, the platform will initially utilize Google Vertex AI, allowing for easy switching between different cloud-based language models, thus ensuring greater flexibility and scalabilityAttribution-NonCommercial 3.0 Brazilhttp://creativecommons.org/licenses/by-nc/3.0/br/inteligência artificial generativaAcesso à informaçãoRAGChatbotAssistente virtualbachelorThesis