Mercado Libre / Mercado Pago: Leading Latin American commerce and digital-payments ecosystem

Problem

Control fraud and chargeback exposure without creating excessive false declines or constraining legitimate payment growth.

Role

Data Science Supervisor contributing to transaction-risk scoring, fraud prevention, and payment decisioning.

Approach

Contributed to real-time transaction-risk scoring, cutoff optimization, behavioural and velocity signals, buyer and merchant risk assessment, and prioritization of cases for manual review.

Impact

Contributed to fraud-risk models and payment decisioning that helped protect Mercado Pago’s growing transaction volume. In 2017, approximately US$9.6 billion in marketplace payments were processed through Mercado Pago, representing 81.9% of Mercado Libre’s GMV, while the company formally reported using anti-fraud models to reduce payment losses.

Lessons

Fraud prevention is a multi-objective decision: losses, chargebacks, approval rates, false declines, review capacity, and customer experience must be optimized together.

Employers are named and consulting clients stay anonymized. Company-level figures such as portfolio size, GMV or delinquency show the context of the period in which the work took place.