I've interviewed at about 15 companies for AI roles.

Some were great. Some had red flags visible from orbit.

Red flags I wish I'd recognized earlier:

"We use AI for everything." — If they can't tell you specifically what problem AI solves, they're in hype-driven development mode.

"The model is already built, we just need someone to deploy it." — Translation: a data scientist built a notebook prototype and nobody knows how to make it production-ready. You'll spend months untangling.

No MLOps infrastructure at all. — If they're manually deploying models, you're signing up for firefighting, not engineering.

"We need someone to build our entire AI platform. Starting next week." — One person cannot build an AI platform. This is a team job staffed as a solo role.

"Our data is perfect, we just need the model." — Nobody's data is perfect. If they believe this, they haven't looked closely enough.

Green flags: clear problem definition, existing data infrastructure (even basic), realistic timelines, a team with complementary skills, and leadership that understands ML is iterative.

The best AI roles aren't at the most famous companies. They're at companies where AI solves a clear problem and the team understands what it takes to build it well.