Fernando Maia da Mota

Faculty of Computing (FACOM)

Federal University of Mato Grosso do Sul

Campo Grande, MS, Brazil

I am a PhD student in Computer Science at the Faculty of Computing (FACOM) of the Federal University of Mato Grosso do Sul (UFMS), Brazil, advised by Prof. Edson Takashi Matsubara and co-advised by Prof. Jonathan de Andrade Silva.

Research. My doctoral research is on text anonymization under the inference capabilities of large language models. Removing names and other explicit identifiers from a document is no longer enough: a language model can often work out who a person is, or sensitive facts about them, from the context that remains. I study how to measure this re-identification risk and how to anonymize documents so that they stay useful for the tasks they are shared for. My application domain is legal documents in Brazilian Portuguese.

Background. Before the PhD, I spent more than a decade building data platforms in industry: data warehouses, ETL/ELT pipelines, data lakehouses and analytics for agribusiness, construction and the public sector. I still work on tools for that world, such as PBI Grid, a dashboard-as-code compiler for Power BI, and Rampart, a framework for reproducible benchmarks of data-processing engines.

I hold a BSc in Information Systems (2011) and a professional MSc in Applied Computing (2016), both from UFMS. My master’s work combined OLAP and data mining to support decision making in beef cattle production. As an undergraduate, I worked on publishing statistical data as Linked Data with researchers from PUC-Rio and the University of Leipzig.

Interests: privacy in natural language processing, large language models, information extraction, knowledge graphs and data engineering.

news

Sep 08, 2026	Our paper on PBI Grid, a dashboard-as-code tool for Power BI reports that follow the GovBR standard, was published at SBES 2026 in São Paulo.
Sep 08, 2026	Our paper on Rampart, reproducible benchmarks of data-processing paradigms with automated anti-leakage checks, was published at SBBD 2026.

selected publications

SBES
PBI Grid: A Dashboard as Code Tool for Enhancing Maintainability and GovBR Compliance in Power BI Reports

Fernando Maia Mota, Eos Xavier, Debora Paiva, and 2 more authors

In Simpósio Brasileiro de Engenharia de Software (SBES), 2026

Abs Bib PDF Supp Video Code

Power BI is widely used by Brazilian federal agencies to publish public-sector dashboards. Despite the existence of the Brazilian Government Digital Standard (GovBR), many of these dashboards do not comply with its visual identity, accessibility, and interface consistency requirements. The Power BI Enhanced Report Format (PBIR) exposes pages and visuals as JSON files but still relies on absolute pixel coordinates and lacks declarative layout abstractions. This paper presents PBI Grid, an open-source command-line tool that compiles declarative YAML specifications into PBIR artifacts. The tool combines three main elements: a declarative language based on 12-column grids, a deterministic layout engine that resolves declarative specifications into PBIR-compatible coordinates, and an extensible theming system with a reference govbr package that codifies the GovBR visual identity for Power BI dashboards. PBI Grid was used to reconstruct a production CAPES dashboard under the GovBR standard; on a bundled, redistributable synthetic example (countries_population), it reduces the layout specification by approximately 89% while remaining amenable to code review and continuous integration. An extract command enables incremental adoption by reconstructing YAML specifications from existing PBIR reports.
@inproceedings{mota2026pbigrid, title = {{PBI Grid}: A Dashboard as Code Tool for Enhancing Maintainability and {GovBR} Compliance in {Power BI} Reports}, author = {da Mota, Fernando Maia and Xavier, Eos and Paiva, Debora and Marcacini, Ricardo and Turine, Marcelo}, booktitle = {Simpósio Brasileiro de Engenharia de Software (SBES)}, address = {São Paulo, Brazil}, year = {2026}, }
SBBD
Rampart: Reproducible Benchmarking of Data Processing Paradigms with Automated Anti-Leakage Verification

Eos Xavier, Lívia R. Duarte Ripardo, Fernando Maia Mota, and 2 more authors

In Simpósio Brasileiro de Banco de Dados (SBBD), 2026

Abs DOI Bib HTML PDF

The growing adoption of data-processing engines for machine-learning pipelines has increased the need for rigorous benchmarks that ensure validity and reproducibility. However, many studies still lack mechanisms to prevent data leakage. Rampart is presented as an open-source framework executing the same walk-forward validation pipeline across three data-processing paradigms while enforcing anti-leakage invariants. Rampart treats prediction equivalence as a validity clause for cross-paradigm latency comparisons and produces bitwise-reproducible executions. Evaluated on two public time-series panels, the framework reveals a scale-dependent latency crossover and supports inclusion of new paradigms with low engineering overhead.
@inproceedings{xavier2026rampart, title = {Rampart: Reproducible Benchmarking of Data Processing Paradigms with Automated Anti-Leakage Verification}, author = {Xavier, Eos and Ripardo, Lívia R. Duarte and da Mota, Fernando Maia and Nogueira, Bruno M. and Borges, Vanessa A.}, booktitle = {Simpósio Brasileiro de Banco de Dados (SBBD)}, publisher = {SBC}, pages = {728--734}, year = {2026}, doi = {10.5753/sbbd.2026.249291}, }

IJSC

Publishing Statistical Data on the Web

Percy E. Rivera Salas, Fernando Maia Mota, Karin K. Breitman, and 3 more authors

International Journal of Semantic Computing, 2012

Bib

@article{salas2012ijsc,
  title = {Publishing Statistical Data on the Web},
  author = {Salas, Percy E. Rivera and da Mota, Fernando Maia and Breitman, Karin K. and Casanova, Marco A. and Martin, Michael and Auer, Sören},
  journal = {International Journal of Semantic Computing},
  volume = {6},
  pages = {373--388},
  year = {2012}
}