Data Scientist
Current1. Developed DWS feature-wide tables based on different business processes in the data warehouse (DWD facts table) and aggregated them into basic features for recommendation scenarios in multiple dimensions.2. Utilized Spark, Shell, Python, and other tools for data processing, such as binning and normalization, enriching model features.3. Developed and optimized live streaming CTR and post-validation (danmaku/likes/payments) models, achieving over 35% improvement in relevant metrics.4. Developed and optimized push notification strategies, resulting in a 30%+ increase in click-through rates and a 40%+ increase in room entry UV.5. Developed and maintained log and monitoring data, including ETL, analysis, and alerts. Utilized technologies such as Kafka, Flink, Spark, and Superset.