Applied AI and machine learning within software systems architecture: a dozen AI, machine-learning and data-mining projects since 2019, from the Sinntesis tender-analysis tool at the São Paulo Court of Accounts to BeyondRisk's LLM analytics, Lionsgate's recommendation APIs and the Claim.md AI/ML programme.
Explore My AI WorkMy journey
Software systems architect by training (D.Sc., COPPE/UFRJ). My field is software systems architecture and applied computing — distributed, data-intensive and AI-enabled systems — across 26 years, 168 projects and 66 organisations in Brazil and the United States.
Since 2019 my practice has centred on applied AI and machine learning: Postgraduate Diploma in Machine Learning and AI (Emeritus with Columbia Engineering Executive Education, 2020, grade 9.22/10) and MIT Sloan Executive Education's blockchain programme (2019). A dozen AI, machine-learning and data-mining projects followed — from the Sinntesis tender-analysis tool deployed at the São Paulo Court of Accounts to BeyondRisk's data lake and LLM analytics, Lionsgate's generative-AI recommendation APIs and the Claim.md AI/ML programme.
Editor of IGI Global's 2026 volume on large language models for business analytics, and Editorial Review Board member of two IGI Global journals, with 32 manuscript reviews — several on AI and game-based learning systems.
Areas of expertise
Hands-on experience across the full AI and ML spectrum — from theory to production.
Extensive experience integrating and fine-tuning LLMs for business applications. Proficient with GPT, BERT, and derivatives to create intelligent, context-aware solutions that enhance user experiences and operational efficiency.
Advanced NLP for sentiment analysis, document classification, text summarization, and multilingual applications. Bridging human communication and computational understanding in production systems.
Developing predictive models that forecast trends and provide actionable insights. Experience with time series analysis, regression models, and ensemble methods for finance, EdTech, and customer behavior.
Designing robust, scalable AI architectures on AWS. Creating efficient data pipelines, optimizing model deployment, and integrating AI capabilities seamlessly into existing enterprise systems at scale.
Developing sophisticated conversational interfaces that understand context, maintain conversation flow, and deliver value through natural interactions — from rule-based systems to advanced neural network approaches.
Designed and implemented recommendation engines using collaborative filtering, content-based filtering, and hybrid approaches. Significantly improved user engagement for Lionsgate and Newsela.
Featured work
Production AI systems that delivered measurable business impact.
Core engine of BeyondRisk's AI application builder and its Prompt to Build front end: domain-specific languages and LLM APIs that turn specifications into deployable applications. Part of the data platform architected as Head of Technology.
Data lake on RabbitMQ, Apache Iceberg, Spark and Trino, with AI agents and LLM APIs that turn cybersecurity data into executive-ready business-intelligence insights. Built with a team of about ten engineers; MVP in four months.
Generative-AI title-recommendation APIs for Lionsgate's streaming platforms, via BairesDev: viewing patterns and content metadata turned into personalised recommendations at production scale.
Advanced AI for financial market analysis and prediction. ML algorithms and time series analysis identify patterns and opportunities in financial data streams, providing traders with actionable insights.
Machine-vision board that watches a physical Kanban board through a camera (OpenCV, TensorFlow) and tracks task movements across workflow stages — the first in Brazil — built for Project Builder.
AI and data-mining system for the prior analysis of public-procurement notices, deployed at TCE-SP as the Court's data-mining tool and since migrated into Marzullo Soluções' tool set — the system for which BrazilLAB selected and certified Sinn.
Conversational AI chatbot integrated with the REGIN platform that streamlines company registration. Guides users through complex procedures in natural language.
Microservices for Newsela's content recommendation engine using ML models for content similarity and clusterization, improving educational content discovery for students and teachers.
Lead professional, through MZ Unlimited LLC, for Claim.md's AI/ML, cybersecurity, solutions-architecture and engineering programme since May 2026, directing a delivery team of eight to nine engineers engaged as independent contractors.
MZ Unlimited's own product line: modular business applications over a shared data layer with AI-driven workflows. The first module, Ledgi Start, is in internal use; further modules are being developed toward commercial release.
The twelve AI-related entries of the 168-project register, with the engaging organisation and the years where recorded.
Continuous learning
The credentials behind the AI practice, as filed: a doctorate in systems and computer engineering, the Columbia/Emeritus diploma and MIT Sloan's blockchain programme. Databricks and DataCamp courses complete the picture.
Emeritus with Columbia Engineering Executive Education
Overall grade 9.22/10 (Applied Machine Learning 9.29, Applied Artificial Intelligence 9.15); capstone passed. Followed by Python for Data Science (Columbia/Emeritus, August 2020).
MIT Sloan School of Management Executive Education (with GetSmarter)
Executive-education programme on blockchain technologies and their business application; part of the 2019–2020 specialisation in AI, machine learning and distributed systems.
COPPE / UFRJ
Thesis Serviços de Desenvolvimento de Software: Uma Abordagem para Reutilização de Modelos Conceituais — software-development services: an approach to the reuse of conceptual models — under Professor Jano Moreira de Souza; defended March 2014.
Databricks
Advanced course covering the inner workings of LLMs — from the basics of the transformer block and attention mechanism to the latest advances in multi-modal LLMs.
Datacamp
Fundamentals of unsupervised learning and implementation of essential algorithms using scikit-learn and SciPy, covering clustering, dimensionality reduction, and anomaly detection.
Research & thought leadership
Two edited IGI Global volumes, eighteen peer-reviewed papers (2006–2011), 94 Google Scholar citations and editorial service for two IGI Global journals.
IGI Global · published 20 May 2026
Edited volume, edited by Fabio Marzullo and published on 20 May 2026, exploring practical applications of large language models in business intelligence and data analytics: implementation strategies, case studies and best practices for extracting actionable insights from enterprise data. ISBN 9798369366905 (e-book 9798369366929) · DOI 10.4018/979-8-3693-6690-5 · Library of Congress cataloguing names Fabio Perez Marzullo as editor.
IGI Global, 2021 · indexed in Scopus and the Web of Science Book Citation Index
Co-edited with Felipe Antonio de Oliveira: eight chapters by authors in seven countries (Brazil, Portugal, Oman, Greece, the United Kingdom, Latvia, Ethiopia). Foothill College (California) lists it among the three representative texts of its course LINC 60C — Educational Game Design; it is held or recorded by the University of New Brunswick, the University of Southampton repository, Afyon Kocatepe University, the Taiwan Academic E-Book Consortium catalogue, Kuwait University and others, and its chapters are cited in four 2025–2026 journal articles. DOI 10.4018/978-1-7998-5021-2.
Barcelona · Berlin · Portland · Honolulu · Las Vegas · Rome · Vancouver · Paris · Cologne
Nine academic events in four countries: ICEIS 2008 Barcelona · ECMDA-FA 2008 Berlin · IEEE/ACIS ICIS 2008 Portland · IEEE SERVICES / SOA Industry Summit 2008 Honolulu · WORLDCOMP 2006, 2008, 2009 and 2010 Las Vegas · IADIS WWW/Internet 2009 Rome. Industry events: XDS 2017 and 2019 Vancouver, Game Connection Europe 2017 Paris and Gamescom 2018 Cologne (Arcolabs delegate, presenting Amazing Wise Friends); Anima Forum 2016, Rio (consultant-evaluator, Anima Business: Coaching).
Eighteen articles, chapters and conference papers published by Springer, IGI Global, the Association for Information Systems, IEEE-, IADIS-, INSTICC- and CSREA-organized conferences and Brazil's CONSAD public-management congress, on model-driven architecture, domain-driven development, SOA and IT governance in government. 94 Google Scholar citations to eleven works; the ICIS 2010 paper has 244 downloads and 547 PlumX interactions on the AIS eLibrary, with readers in 37 countries.
IGI Global · since July 2017
Editorial Review Board member of two IGI Global journals — the International Journal of Game-Based Learning and the International Journal of Distance Education Technologies — and of IGI Global's Research Insights Editorial Advisory Board (since July 2017). 32 completed manuscript evaluations (December 2017 – October 2026), several on AI and game-based learning systems; seven ranked "Very Strong" by the editors; 15 reviews deposited at ORCID 0000-0003-1693-9146.
Reach & Impact: adoption, citations and editorial serviceLet's build something
Available for strategic consulting, AI/ML architecture, LLM integration, and technical leadership engagements worldwide.