Spark / Fusion
Performance-oriented Spark platform work, including Fusion-related acceleration and serverless lakehouse analytics.
Alibaba Cloud EMR
Cloud-scale data products.
Product Lines
StarRocks / Stella
OLAP and lakehouse analytics work around EMR StarRocks and Stella Kernel performance leadership.
Milvus
Vector database work for AI-era retrieval, multimodal data, and vector lake scenarios.
EMR Agent
AI-assisted EMR operations and product experience for modern cloud data platforms.
EMR Scope
DataOps Platform
Product and engineering work for cloud data platform development, operations, and user workflows.
Computing Engines
Flink, Hive, Spark, StarRocks, Trino, and related open-source engines in Alibaba Cloud EMR.
Metadata & Security
Hive Metastore, Ranger, and the management layer needed by enterprise data platforms.
Lake Storage
Cache, lake format, lake file system, and data lake storage engines for lakehouse and AI-era workloads.
EMR Evolution
My EMR work has focused on evolving the platform from open-source component management toward cloud-native and serverless data infrastructure, including EMR 2.0, EMR Serverless Spark, lakehouse analytics, and AI-assisted data platform capabilities.
More recently, this direction has expanded into OpenLake, lake-stream integration, multimodal data storage and access, and the shift from human-centric data access patterns to AI-agent-driven data generation and retrieval.
Earlier Systems Work
Earlier systems work in HBase, search data infrastructure, realtime compute storage, and Flink state backends shaped the current product direction: production-first, open-source-compatible, and measured under real traffic.
Engineering Focus
- Serverless big-data platforms
- Lakehouse analytics engines
- Data + AI platform architecture
- Vector databases and multimodal lake
- Remote shuffle and elastic execution
- Production-ready open-source infrastructure