At Ryax Technologies, I led IntelliScale, an ML-based Kubernetes vertical autoscaler that reduced GPU consumption by 31.2% while maintaining workload makespan. I also built NVIDIA MIG-based GPU autoscaling and a secure HPC API prototype connecting Kubernetes and Slurm clusters for CEA.
Previously, I rebuilt Tencent Yangshipin's search Query Understanding platform as Go microservices and developed Kafka, Spark, and HBase systems for large-scale video-search indexing and analytics. My work spans distributed systems, cloud-native platforms, HPC scheduling research, and real-time multiplayer game systems.
