Yaming Zhang

• Core infra & Data Platform

• Designed the Java-based OOP model for Player Lifecycle Graph, a serialization model to manage players’ states and unified messages, resulting in a 30% cost reduction by eliminating the need for customizing Braze
• Developed a Flink topology for Kafka streaming PIN events execution, enabling efficient retrieval of lifecycles from Redis, reading player profiles from ScyllaDB, and executing the corresponding lifecycles
• Leveraged EADP’s Data Query Language (DQL) to extract data from unstructured JSON events, evaluate boolean conditions, and execute action scripts
• Built a web application with React and implemented APIs with Spring Boot, empowering campaign managers to draw lifecycle flowcharts and retrieve players’ current states in games

Tencent News @ PCG

• Built Tencent News Risk Management Platform using tRPC with Redis, Elasticsearch, etc
• Analyzed Tencent News App traffic data using Spark, Hive and Presto to do the OLAP for the team

Tencent Healthcare @ CSIG

• Developed Microservices on Go using gRPC with logs on ELK, monitoring system on Prometheus, Grafana, asynchronous pipelines to Kafka, etc, achieving a 25% improvement in interface response time and a 50% reduction in production debug time
• Established automated CI/CD pipelines that enhanced service security by seamlessly integrating Unit Tests during CI, while also enabling efficient rollback on Tencent Cloud Kubernetes

• Leveraged the power of Scala and Spark to design user-defined functions for geometry calculation, enabling the provision of diverse and flexible APIs that empower developers to generate and manipulate geometries with ease
• Utilized the Hadoop file API to efficiently parse records from the OpenStreetMap PBF file, enabling extraction of OSM entities for subsequent Big Data Processing Pipelines on AWS EMR
• Extracted nodes, POIs, roads and all objects in pipelines and categorized different datasets from all objects based on OSM tags, including parks, lakes and buildings, etc
• Generated the satellite image of corresponding OSM Vector data using Beast Raptor spatial-join for spatial machine learning and AI activities