🍔 Casper's Kitchens

Casper’s Kitchens is a fully Databricks-native ghost kitchen and food-delivery platform built by the Developer Relations team. It brings together every layer of the Databricks platform — Lakeflow (ingestion, Spark Declarative Pipelines), AI & BI dashboards with Genie, Agent Bricks, and Apps powered by Lakebase (Postgres) — into a single, cohesive live demo.

Casper’s is more than a showcase. It’s a living playground for simulation, demos, and creative misuse — designed to push the Databricks platform past its comfort zone.

Everything is built to be easy to:

1. 🚀 Deploy — spin up the entire environment in minutes.
2. 🎬 Demo — run only the stages you need, powered by live streaming data.
3. 🧑‍💻 Develop — extend with new pipelines, agents, or apps effortlessly.

We build only with Databricks — by choice — so Casper’s serves as a shared sandbox for learning, experimentation, and storytelling across the platform.

Prerequisites

• Databricks CLI installed on your local machine.
• Authenticated to your Databricks workspace. (can do interactively databricks auth login)
• Access to the repository containing Casper's Kitchens.
• Permissions in the Databricks workspace to create new catalogs.

🚀 Deploy

Casper’s Kitchens uses Databricks Asset Bundles (DABs) for one-command deployment. Clone this repo, then run from the root directory:

databricks bundle deploy -t <target>

Each target represents a different flavor of Casper’s (for example, full demo, complaints-only, free tier, etc.). Use whichever fits your needs:

databricks bundle deploy -t default     # full version: Data generation, Lakeflow, Agents, Lakebase & Apps
databricks bundle deploy -t complaints  # complaints agent: Data generation, Lakeflow, Agents, Lakebase
databricks bundle deploy -t free        # Databricks Free Edition: Data generation, Lakeflow

This creates the main job Casper’s Initializer, which orchestrates the full ecosystem, and places all assets in your workspace under /Workspace/Users/<[email protected]>/caspers-kitchens-demo.

💡 You can also deploy from the Databricks UI by cloning this repo as a Git-based folder and clicking Deploy Bundle.

For more about how bundles and targets work, see databricks.yml or the Databricks Bundles docs.

🎬 Run the Demo

Once deployed, run any target with the same command:

databricks bundle run caspers

Optionally, specify a catalog (default: caspersdev):

databricks bundle run caspers --params "CATALOG=mycatalog"

This spins up all the components—data generator, pipelines, agents, and apps—based on your selected target.

To clean up:

databricks bundle run cleanup (--params "CATALOG=mycatalog")
databricks bundle destroy

🧩 You can also run individual tasks or stages directly in the Databricks Jobs UI for finer control.

📊 Generated Event Types

The data generator produces the following realistic events for each order in the Volume caspers.simulator.events:

Event	Description	Data Included
order_created	Customer places order	Customer location (lat/lon), delivery address, ordered items with quantities
gk_started	Kitchen begins preparing food	Timestamp when prep begins
gk_finished	Kitchen completes food preparation	Timestamp when food is ready
gk_ready	Order ready for pickup	Timestamp when driver can collect
driver_arrived	Driver arrives at kitchen	Timestamp of driver arrival
driver_picked_up	Driver collects order	Full GPS route to customer, estimated delivery time
driver_ping	Driver location updates during delivery	Current GPS coordinates, delivery progress percentage
delivered	Order delivered to customer	Final delivery location coordinates

Each event includes order ID, sequence number, timestamp, and location context. The system models realistic timing between events based on configurable service times, kitchen capacity, and real road network routing via OpenStreetMap data.

🎯 Use Cases

• 📚 Learning Databricks: Complete end-to-end platform experience
• 🎓 Teaching: Consistent narrative across different Databricks features
• 🧪 CUJ Testing: Run critical user journeys in realistic environment
• 🎨 UX Prototyping: Fully loaded platform for design iteration
• 🎬 Demo Creation: Unified narrative for new feature demonstrations

Check out the Casper's Kitchens Blog!

License

© 2025 Databricks, Inc. All rights reserved. The source in this notebook is provided subject to the Databricks License [https://databricks.com/db-license-source]. All included or referenced third party libraries are subject to the licenses set forth below.

library	description	license	source