Databricks announced major updates to Lakeflow, its unified data engineering platform, at the Data + AI Summit. Key highlights include: Genie Code AI assistant integrated across ingestion, transformation, and orchestration; Lakeflow Designer (now GA) offering a no-code drag-and-drop pipeline builder; Genie ZeroOps, a new background AI agent for automated pipeline monitoring, root-cause analysis, and fix generation; Lakeflow Connect expanding to 100+ native managed connectors with a free tier (100 DBUs/day); Zerobus Ingest for near real-time high-throughput event ingestion (under 5s latency, up to 100MB/s); Real-Time Mode for Spark Declarative Pipelines now in public preview with end-to-end latencies as low as 5ms; and Lakeflow Jobs adding data-readiness triggers and external orchestration with 40+ operator examples.

8m read timeFrom databricks.com
Post cover image
Table of contents
Genie Code and Lakeflow Designer: agentic pipeline developmentGenie ZeroOps: Put data and AI operations on autopilotLakeflow Connect: Fast-growing ecosystem with 100+ built-in connectorsZerobus Ingest: Kafka-free ingestion for your data producersSpark Declarative Pipelines: Batch and streaming, SQL and Python, and now real-timeLakeflow Jobs: Now with 50+ integrationsGetting started with Lakeflow
249 Impressions