At Sindpass, I analyze fare collection data to identify patterns, anomalies, and potential indicators of fraud. I monitor transactions and atypical user behavior, supporting investigations and fraud prevention measures.
On my Smart Ticket Fraud Detection portfolio project, I built an Azure Databricks medallion pipeline that processed 7.77 million fare transactions. I added data quality checks, governance and lineage, and SQL marts with dbt.
I developed fraud detection using Isolation Forest and SHAP, achieving 0.916 ROC-AUC and 97% precision on the audit queue, with an explanation for every alert. I also delivered a four-page Power BI report.
Earlier, at Floresta Supermarket, I monitored inventory and shrinkage KPIs and audited pricing, inventory, and invoicing data. At Youtility Center do Brasil, I recorded customer interactions in CRM and ServiceNow and analyzed ticket history to support prioritization.
