Data · Fraud · Payments — Data Analyst, Fraud Strategy & Data Science
Detecting anomalous patterns at scale — the same analytical muscle as genomic surveillance, pointed at transactions instead of pathogens.

LinkedIn · Lattes · Email

At InComm I work inside the Fraud Strategies team: transaction-level analysis in SQL Server (aggregation, window functions for classification within groups, joins, data-quality checks), fraud detection and prevention strategy, and Power BI dashboards for fraud monitoring.

I came to fraud analytics from genomic disease surveillance. As a CNPq research scholar at NUPI/EBMSP I spent two years finding emerging signals in large population datasets — arboviruses and COVID-19 — with Excel and statistics.

The problem shape hasn't changed, only the data: anomaly and pattern detection at scale.

Member of the Micro Sphere Research (MSR) group since 2025, in bioinformatics and epidemiology.

genai-biomedical-nlp — an end-to-end GenAI + NLP pipeline over PubMed literature (arbovirus surveillance): collection via NCBI Entrez → zero-shot classification + biomedical NER (Hugging Face) → LLM summarization → validation against a hand-labeled holdout. MIT, runs free on Colab.

• variantscribe — agentic clinical variant interpretation: drafts ACMG/AMP classifications with cited evidence, designed to be evaluated against ClinVar gold labels. RAG + LLM eval + LLMOps.
• dual-agent-review — cross-agent code review: Claude Code and Gemini CLI audit each other, the human decides.

_{🇧🇷 Based in Bahia, Brazil · open to remote / relocation}

_{Anomalies, at scale. Always learning.}