"Technology is best when it brings people together."
Matt Mullenweg
Researcher and developer in Artificial Intelligence, working with Natural Language Processing, Computer Vision, Multi-Agent Systems and Multimodal Processing.
- Data Scientist at SETIC, Rondônia State Superintendency of Information Technology
- B.Sc. in Computer Science, Federal University of Rondônia (UNIR), 2025
- Postgraduate student in Applied Artificial Intelligence, Federal University of Paraná (UFPR)
- Special student at ICMC/USP (PPG-CCMC), Research in Computing with Data
- Current research: testing of AI agents, AI-generated image detection, bioinformatics and medical imaging
- Outside of work: Judo black belt
| Area | Focus |
|---|---|
| Artificial Intelligence | NLP, LLMs, AI agents, multi-agent systems, multimodal models |
| Computer Vision | Image enhancement and segmentation, medical imaging (OCT), facial expression recognition |
| Data Science | Graph machine learning, synthetic data generation, MLOps |
| Software Engineering | Full-stack systems with integrated AI modules, metamorphic testing, agile (XP) |
| Role | Organization | Period |
|---|---|---|
| Data Scientist | SETIC, Rondônia State Government | 2026 to present |
| Software Engineer, AI systems, computer vision and image processing | DATAFull Tecnologia | 2025 to 2026 |
| Undergraduate Researcher, OCT image enhancement and segmentation for glaucoma treatment | UNIR / CNPq · Porto Velho Eye Hospital | 2024 to 2025 |
| Teaching Assistant, Computer Organization; Electronics for Computing | UNIR | 2023 to 2024 |
| Student Representative, Computer Science Department Council | UNIR | 2024 to present |
| Samsung N.A.V.E Tech program | Samsung Brasil / UNIR | 2023 to 2024 |
Conference papers
- When the Trigger Fails: Metamorphic Testing of Tool-Calling AI Agents. SAST 2026. Artifact badges: Available, Functional.
- Procedural Framework for Synthetic Data Generation Based on Statistical Simulation: An Approach for Human Behavior Analysis. WebMedia 2025.
- Engineering a Secure and Traceable System for Diploma Issuance in a Brazilian Federal Institution. WSIS 2025.
Accepted (2026)
- SPIN-Seq: Predicting the Typed Residue Interaction Network from Sequence. Brazilian Symposium on Bioinformatics (BSB).
- ProgSinais: Programming Feedback in Libras with LLMs and Verified Signability. WebMedia 2026, Extended Proceedings.
- TokenAudit: Deterministic Verification of Disclosure-to-Enforcement Consistency in ERC-3643 Tokens with LLM-Based Extraction. WebMedia 2026, Extended Proceedings.
Book chapter
- PVPI: A Deterministic Platform for Authenticity Verification, Traceability and Detection of AI-Generated Images. In Perspectivas em Transformação Digital, UFSC, 2025.
| Year | Recognition |
|---|---|
| 2026 | Research Artifact Badges Available and Functional, SAST / CBSoft |
| 2025 | Best Final Report in Undergraduate Research, UNIR / CNPq |
| 2024 | 1st Place, OPEN DAY ADS Programming Marathon, IFRO |
| 2023 to 2024 | Competitor, SBC Programming Marathon |
- Introduction to Machine Learning on Graphs, short course, 2025
- LaTeX for Scientific Writing, short course, 2025
- Introduction to Machine Learning, short course, 2024
- Minecraft as a Platform for AGI Studies: Multimodal Architectures with Persistent Memory, seminar, 2024
- Humanoid robot NAO V6 demonstration, UNIR Open Doors, 2024
