Educational technology leader · Researcher · Builder
Where cognition meets code — and prototypes are expected to survive contact with reality.
I work at the intersection of education, artificial intelligence, and digital systems — turning emerging technology into infrastructure, learning experiences, and research that people can actually use.
At Sesc's National Department, I coordinate educational technology initiatives designed to work across Brazil's very different realities. At PPGI/UFRJ, I research explainable and auditable AI. I also teach, speak, and build tools that make complex ideas easier to understand without pretending they were simple all along.
Based in Rio de Janeiro, I spend a fair amount of time translating between classrooms, institutions, research papers, and codebases. Somehow, they all have edge cases.
| Lead | Research | Build & teach |
|---|---|---|
| National-scale educational technology, digital transformation, platforms, and institutional adoption. | Explainable AI, provenance, human oversight, and systems whose decisions can be inspected. | Open-source products, learning experiences, and technical ideas made useful beyond the lab. |
A deliberately short list. A profile should offer evidence, not require an archaeological expedition through every repository.
| Project | What it does |
|---|---|
| GreenRisk · paper & DOI | An explainable-by-construction instrument for scoring greenwashing risk in corporate climate disclosures, with fuzzy rules, activation traces, and W3C PROV-O provenance. |
| Padline · live | URL-first, no-account collaborative pads with real-time editing, offline resilience, history, and privacy controls — because “open a link and type” is still a very good interface. |
| The Membrane Palette · showcase | A token-based design system where human warmth meets digital precision, built for both human teams and AI agents. |
| Remove AI Writing Signs | A reconstruction-based editing system for making AI-assisted prose specific, uneven, and genuinely human — not merely allergic to the word “delve.” |
- How can AI systems show their reasoning, provenance, and limits without becoming unusable?
- How can educational technology operate at national scale without flattening local reality?
- How do we design for attention, accessibility, and agency from the architecture onward?
- What deserves automation — and what still deserves a person?
Product & interfaces
AI & data
Systems
Tools change. Principles age more slowly.
- Human-centered is an architectural choice, not a decorative adjective.
- Explainability beats theater. If a system matters, its decisions should be open to inspection.
- Small friction compounds. Good tools respect attention, context, and the people operating them.
- Rigor and wit can coexist. I have reviewed the evidence.
I am always interested in thoughtful conversations about responsible AI, educational technology, learning infrastructure, open-source tools, and research meant to survive outside the PDF.
