At NVIDIA, we accelerate Apache Spark with an open source plugin that speeds up Spark applications on GPUs without code changes, and our cuDF Spark library is integrated with AWS EMR, Databricks, Google Dataproc, Oracle Cloud Data Flow, Bytedance Volcengine, Tencent Cloud and Cloudera. Design, build and operate hosted Spark service control plane on Kubernetes: the Java REST API, the Spark Connect runner, Kyuubi, and deployment using Helm charts and Flux.