I’ve been a casual developer across many languages for ~10ish years but got a degree in mechanical engineering. I pivoted and started an analytics engineering job around 3 years ago and starting doing web development for side projects.

Lately my team has leaned heavily into AI-assisted programming primarily using Claude Code to build React Apps in Databricks. These are primarily CRUD apps but sometimes they do write back to our ERP so they affect the business greatly. The AI writes most of the code now, and I’m trying to figure out where my review time is best spent.

Options I’m weighing:

• Full line-by-line review (quality, readability, structure)
• Mainly reviewing the tests, and trusting the code if the tests are solid
• Testing through the UI, then checking the DB/tables to confirm data was written correctly

Does code readability still matter if AI is the one maintaining it? What has bitten you when you skipped review? If you have a concrete review checklist, I’d love to see it.