At Tongdun Technology, I led the architecture and evolution of a One-Stop Big Data Platform, Financial Risk Feature Store, and DataBI Platform supporting 10,000+ daily core jobs. I delivered Spark on Kubernetes with JuiceFS and S3/OSS storage, completed three cross-data-center migrations with zero downtime or data loss, and upgraded enterprise workloads to Spark 3.x.
I build real-time and offline risk-control data systems across Spark, Flink, HBase, Cassandra, and ScyllaDB. My feature lifecycle work shortened new risk feature delivery by 40%, while distributed caching, RPC tuning, and storage redesign achieved sub-20ms retrieval under high concurrency.
