At NOMINIS, I led data science and intelligence-engineering projects supporting the discovery of crypto wallets and wallet-owner attributions for financial-crime investigations. I also partnered with intelligence analysts on selected extraction workflows.
I developed workflows to classify and contextualize text from Telegram, X/Twitter, OCR-derived content, and legal documents, connecting off-chain signals with blockchain entities. I fine-tuned DistilBERT for domain-specific text classification and developed a FastAPI wallet-risk application with LLM evidence grounding and audit logging.
My work also included multi-chain wallet-enrichment pipelines and threat-intelligence records for exploit, phishing, and rug-pull reports. I hold a BSc in Statistics from Federal University of Minas Gerais (UFMG) and co-authored research published in Inflammopharmacology.
