Prompt to Pipeline: Hands-On Data Engineering with Cortex Code

You’ve tried an AI coding agent, and it wrote you some SQL. But did the data actually land?

Data engineering work rhymes. Every day brings the same kinds of pipeline debugging, migration grunt work, and environment checks, never quite identical but never truly new. Generic AI coding agents can write the SQL, but they can’t tell you whether the data arrived, what a change will break downstream, or why your warehouse is spilling. They see your code. They don’t see your data.

So you’re stuck re-explaining your environment every session, double-checking every result by hand, and treating AI as a one-off chat window instead of a teammate that knows how your team works.

Join this 2 hour virtual hands-on lab to build a reusable harness on top of Cortex Code (CoCo), the coding agent connected to your Snowflake environment. CoCo doesn’t stop at writing code. It runs the code, checks whether the data arrived, reads lineage to see what a change will break, and pulls metering to show what it costs. You’ll turn it into your team’s own data engineering assistant, one that learns your conventions, remembers your environment, and stops asking the same questions every session. This lab is built for data engineers and platform teams who are ready to move past one-off AI sessions.

In this session, you will learn:

• The vocabulary and concepts for working well with CoCo: surfaces, the harness, skills, plugins, hooks, and context
• How a real day as a data engineer, from morning pipeline checks to code reviews to migration work, maps to CoCo capabilities you’ll build hands-on
• How to write an AGENTS.md file that captures your team’s conventions, so you never re-explain your environment
• How to build a custom skill that automates a repeatable data engineering workflow, and package it into a plugin with safety hooks that protect you when you’re running three sessions in parallel
• How to build an environment-check skill that reads your own Snowflake account and shows what’s broken, stale, or costing you money
• How to use CoCo across the full pipeline: ingestion with Openflow, transformation with Dynamic Tables, and downstream analytics with semantic views
• A framework for growing the harness over time, adding skills and trusting the agent with more of the routine work, so you can spend your time on the decisions that need a human

You’ll leave with a working assistant you can use Monday morning.