At Coinbase, I analyzed payments-risk data to identify fraud patterns, false positives, and high-risk customer segments. I identified and mitigated ACH fraud involving fraudulent documentation, generating $600K+ in savings.
I also resolved vendor billing discrepancies that saved $1M+ and helped return the Chargebacks system to production, contributing to $3M in annual chargeback savings. I built and maintained Looker dashboards to monitor vendor performance, risk data, and operational metrics.
At Fundação Getúlio Vargas (FGV), I automated customer segmentation and mailing workflows using R and IBM SPSS Modeler, and used SQL and Tableau Server to support digital KPIs and management reporting. Earlier, I developed fraud analytics models at Ipiranga and implemented one that identified approximately R$200K in fraud and operational losses.

