How Delivery Hero ensures compliance & reduces fraud with
summary
Delivery Hero builds a selfie-based identity verification flow using Amazon Rekognition and S3/RDS/SNS to automate rider checks, launches it to production in 6 weeks, and runs it across dozens of markets processing ~250,000 images daily. The automated pipeline cuts photos without faces from ~80% to 0.1%, halves mismatch rates (4% to 2.52%), prevents ~€430k/month in compliance fines, and delivers estimated fraud savings of ~€250k/month, reducing fraudulent activity by ~70%.
AI-written summary · about
chapters (9)
- 0:00Introduction and overview
- 7:09Delivery Hero’s growth and challenges
- 11:42Account fraud problem discovery
- 15:57Requirements and UAE legislation impact
- 20:26Amazon Rekognition solution overview
- 22:14Technical architecture deep dive
- 27:16Code implementation insights
- 31:31Results and compliance success
- 33:15Future improvements and Q&A
related talks
29mCOVID-19: Identity verification & work safety with Amazon Rekognition
38mUsing Artificial Intelligence to Verify Identity
36mInformation Reliability Service - a novel approach to fraud detection
14mTransforming Digital Identity with AI
39mFraud Detection: Using ML to Identify and Manage Fraudulent Activities
14mNext-Gen Fraud Strategies Use Data to Onboard Customers Safely
27mEnsuring biometric recognition is secure and works well for all: A Case Study
28mML solutions for the real world problems of identity verification