Yuer Zhou

Master's

Environmental Engineering

Meituan

Machine Learning Engineer Intern

- China - Ads - Implemented incremental training on a vast user base(10M-100M DAUs) for 24-hour and 7-day CVR prediction, transitioned - Developed and evaluated models like XGBoost, Wide&Deep, MMoE, and ESMM using TensorFlow. Implemented data pipeline - Deployed the final optimized MMoE model integrated with delayed feedback modeling and transfer learning techniques - Achieved +1.8pp o✏ine AUC improvement - 5% customer acquisition, -9.3% CAC reduction in online A/B tests - Recommendation - Established a seamless Click-to-Landing recommendation system in Java for users brought in via external ads, including end-to-end - Optimized landing page ranking models by developing 50+ additional key features, utilizing Customized Gate Control(CGC) model