Ciência de dados 2.0 arquitetura computacional e orquestração de sistemas agênticos em ambiente fintech

##plugins.themes.bootstrap3.article.main##

Sandra de Oliveira Soares Cardoso
José Augusto Theodósio Pazetti

Resumo

A consolidação de modelos de linguagem de grande escala (LLMs) e sistemas baseados em agentes autônomos redefine os fundamentos operacionais da Ciência de Dados contemporânea. Este estudo investiga a emergência da denominada Ciência de Dados 2.0 como paradigma arquitetural orientado à orquestração computacional, governança distribuída e otimização econômica de sistemas inteligentes. A pesquisa foi conduzida por meio de revisão sistemática estruturada segundo o protocolo PRISMA, contemplando publicações indexadas entre 2023 e 2026 nas bases IEEE Xplore, ACM Digital Library, Scopus e arXiv. Os resultados evidenciam a transição de pipelines batch monolíticos para arquiteturas distribuídas baseadas em Data Mesh, Lakehouse transacional (Apache Iceberg), bancos vetoriais, Retrieval-Augmented Generation (RAG), Feature Stores em tempo real e práticas de FinOps. Propõe-se o Framework O³ (Orquestração, Observabilidade e Otimização) como modelo integrador capaz de articular desempenho algorítmico, eficiência econômica, rastreabilidade e conformidade regulatória. A validação em estudo de caso aplicado em ambiente fintech orientado a crédito digital demonstra redução significativa de latência, otimização de custo por inferência e aumento da robustez decisional. Conclui-se que a Ciência de Dados 2.0 configura-se como disciplina arquitetural sistêmica, superando a abordagem centrada exclusivamente em modelagem estatística.

Downloads

Não há dados estatísticos.

##plugins.themes.bootstrap3.article.details##

Como Citar
CARDOSO, S. de O. S.; PAZETTI, J. A. T. Ciência de dados 2.0: arquitetura computacional e orquestração de sistemas agênticos em ambiente fintech. Revista Datapoint, [S. l.], v. 1, n. 01, p. 73–88, 2025. DOI: 10.5281/zenodo.19240921. Disponível em: https://www.fatecrl.edu.br/revista/datapoint/index.php/dp/article/view/6. Acesso em: 19 maio. 2026.
Seção
Artigos

Referências

BOMMASANI, R. et al. On the opportunities and risks of foundation models. Stanford University, 2021.

BRASIL. Lei n° 13.709, de 14 de agosto de 2018. Lei Geral de Proteção de Dados Pessoais (LGPD). Brasília, DF: Presidência da República, 2018.

CAVOUKIAN, A. Privacy by design: the 7 foundational principles. Information and Privacy Commissioner of Ontario, 2011.

DEHGHANI, Z. Data Mesh: Delivering Data-Driven Value at Scale. Sebastopol: O’Reilly Media, 2022.

DOSHI-VELEZ, F.; KIM, B. Towards a rigorous science of interpretable machine learning. arXiv, 2017.

EUROPEAN COMMISSION / EUROPEAN PARLIAMENT. Regulation on Artificial Intelligence (AI Act). Brussels: European Union, 2024.

FINOPS FOUNDATION. The FinOps Framework. 2023.

FLORIDI, L. et al. AI4People—An ethical framework for a good AI society: opportunities, risks, principles, and recommendations. Minds and Machines, v. 28, p. 689–707, 2018.

KLEPPMANN, M. Designing Data-Intensive Applications: The Big Ideas Behind Reliable, Scalable, and Maintainable Systems. Sebastopol: O’Reilly Media, 2017.

LEWIS, M. et al. Retrieval-augmented generation for knowledge-intensive NLP tasks. Proceedings of NeurIPS, 2020.

LIBERATI, A. et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions. BMJ, v. 339, b2700, 2009.

MOHER, D. et al. Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Medicine, v. 6, n. 7, e1000097, 2009.

PAGE, M.J. et al. PRISMA 2020 explanation and elaboration: updated guidance and exemplars for reporting systematic reviews. BMJ, v. 372, n. 71, 2021.

PETTICREW, M.; ROBERTS, H. Systematic Reviews in the Social Sciences: A Practical Guide. Malden: Wiley-Blackwell, 2006.

REIS, J.; HOUSLEY, M. Fundamentals of Data Engineering: Plan and Build Robust Data Systems. Sebastopol: O’Reilly Media, 2024.

SIGELMAN, B.; BARROS, A. et al. Dapper, a large-scale distributed systems tracing infrastructure. Google Research, 2010.

STRUBELL, E.; GANESH, A.; MCCALLUM, A. Energy and policy considerations for deep learning in NLP. ACL, 2019.

WOOLDRIDGE, M. An Introduction to MultiAgent Systems. Wiley, 2009.

YAO, S.; BUTTERWICK, J.; et al. ReAct: Synergizing reasoning and acting in language models. ACM Transactions on Intelligent Systems, 2023.

ZHU, X. et al. GraphRAG: Integrating knowledge graphs with retrieval-augmented generation. Journal of Artificial Intelligence Research, 2024.

WORLD ECONOMIC FORUM. Future of Jobs Report 2025. Geneva: WEF, 2025.