Ultra performant compute framework for AI workloads, with incremental processing and data lineage out-of-box.