Build the same Databricks job with Claude Code and Cursor

Watch Claude Code and Cursor each build a Databricks job that pulls monthly Arizona weather data from a free API.

Who it's for: Data engineers and analysts choosing between Claude Code and Cursor.

The problem: With so many AI coding tools, it can be hard to know what to use or how they differ. Claude Code reads CLAUDE.md for context, while Cursor reads AGENTS.md.

Cursor first turns the existing Claude Code context into its own rules, commands and agents. Then both tools get the same prompt: create a Jira ticket, research free climate data APIs, deploy a monthly Databricks job, test it and close the ticket. Both jobs succeeded and ran in under a minute.

Takeaways:

  • Choose one tool and dive in, since the principles carry over.
  • The more specific your context files, the better your results.

Resources