Jonathan Lumbantobing

I've analysed transaction, campaign, and operations data at Shopee Indonesia and Kredivo Group, helping teams turn large-scale data into practical decisions.

At Kredivo, I identified concentrated merchant cashout fraud reaching approximately 98M IDR at the worst site in one month and helped build an early-detection trigger that shut down most high-risk sites. I analysed millions of daily transaction records using SQL and Python while working with Fraud Ops, Product, Data Science, and Engineering to improve rule-based detection.

At Shopee, I built the SQL reports and dashboards used by payments, transactions, customer service, and refund operations teams. I improved report automation to reduce failures and wait times by up to 50%, contributed to recovering misused vouchers worth up to USD 10,000, and used A/B testing to improve banner click-through rates by up to 10%.

I'm completing a Master of Business Analytics at Melbourne Business School, where I focus on causal inference, predictive modelling, machine learning, and analytics for business. In a 14M-row Criteo advertising experiment, I built uplift models and targeting policies that showed the top 5% of an audience delivered 43% of a campaign's total effect.