Resources and insights
Our Blog
Explore insights and practical tips on mastering Databricks Data Intelligence Platform and the full spectrum of today's modern data ecosystem.
Lakeflow Jobs: From Schedules to Events and Conditions
A schedule can’t tell when your source is late. Lakeflow Jobs can now start on a file arrival, a table update, a finished job or a new model version, and a SQL condition checks the work is actually ready.
Genie Ontology PART 2: A Practical Walkthrough in Databricks
Step-by-step guide to Genie Ontology in Databricks: document a table, create a metric view, build a domain, and certify a definition for Genie.
AI_DECIDE() vs 20 LLMs
Databricks AI_DECIDE() is free in beta, so how does it compare? A post moderation test against 20 LLMs measured accuracy, speed and cost, and AI_DECIDE answered all 15 correctly with the fastest median response.
Genie Ontology PART 1: Build Business Context for AI
How Genie Ontology gives Databricks Genie business context: discovering knowledge, OntoRank, governed definitions, and evaluating answers.
From Model to Production: Building a Governed ML Lifecycle on Databricks
Most ML teams can train a model, but far fewer can operate one. This post shows how to build a governed ML lifecycle on Databricks, with automated validation, champion-challenger promotion, one-swap rollback and continuous monitoring, using MLflow and Unity Catalog.
Environment Variables in Databricks Serverless
Copying the same job for dev, staging, and prod invites drift and leaked secrets. Here’s how serverless environment variables let one job definition carry different settings per task, with a working three-task demo.
Automatic Change Data Feed (CDF) is here to save space, improve write time, and feed Change Data Capture (CDC)
Classic change data feed saves an extra copy of every change, which adds storage and slows updates. Hubert Dudek benchmarks the automatic alternative on a 5-million-row table and walks through how it works.
Beyond Bronze, Silver, and Gold: Designing Medallion Layers Around Real Data Needs
Bronze, Silver, and Gold are a great starting point, but real projects rename layers, split them, and stack them. Here’s how to decide what each layer is responsible for, and when a separate one earns its place.
Databricks DABs CI/CD: 4 Ways to Automate Your Deployments
Every CI/CD pipeline eventually needs something to run after deployment: initial setup, config syncs, validation checks. job_runs in Databricks Asset Bundles handles all of it natively, and you can use it to block a production update when validation fails.
Automated Tags Clean The Mess
Learn how to use Databricks Unity Catalog's governed tags to automatically flag undocumented or inactive tables, notify owners, and build a self-governing data catalog.
Databricks as a Model Provider: The Safe Route into Enterprise Adoption of Open-Weight Models
Nicanor Medina (Databricks MVP) explains how Databricks is emerging as a top-tier inference provider for open-weight models like Kimi K3 and GLM-5.3. Fast, affordable, and enterprise-secure.
Genie ZeroOps: A Hands-On Preview
A hands-on preview of Genie ZeroOps, Databricks' autonomous agent for monitoring and fixing data pipeline issues — tested against a data overflow bug, a division-by-zero error, and a live migration.
DABs direct mode: available, faster, and better.
Databricks Asset Bundles direct mode is now GA, and it's already the default for bundles created from the workspace, with full CLI default rollout on August 26, 2026. By removing the Terraform dependency, direct mode cuts deployment time by up to 40%, replaces opaque state with human-readable JSON, and unlocks resources like instance pools and Genie Agents that Terraform-gated releases used to delay.
Databricks Zerobus vs. Kafka: Do you still need Kafka in front of your Lakehouse?
Does 'streaming means Kafka' still hold for single-sink ingestion into Delta? We benchmarked it with 64.8 million real events and found a cost gap that never closes, 3.6× to 20.9×
I benchmarked the new Lakehouse RT: It’s fast, cheap, and open-format
Most teams assume faster real-time queries mean higher compute costs. We benchmarked Databricks' new RT Lakehouse engine across datasets from 1 million to 50 billion records to find out, and the results reset what "fast" costs on a lakehouse.