AI Isn't Taking Jobs — It's Creating Them

The categories that didn't exist before

Scan any AI-heavy company's job board and you'll find titles that were vanishingly rare in 2022.

• LLM / applied AI engineer — building retrieval, tool use, and agent systems on top of foundation models.
• Evaluation and red-team specialist — designing benchmarks and adversarial tests for model behavior.
• AI safety and policy — alignment research, misuse mitigation, and deployment review.
• MLOps and inference infrastructure — serving models cheaply and reliably at scale.
• AI product manager — deciding what a probabilistic system should and shouldn't do.
• Forward-deployed and solutions engineer — adapting models to a specific customer's data and workflow.

What actually changes inside existing roles

Most jobs aren't disappearing; the task mix inside them is shifting. Engineers spend less time on boilerplate and more on review and system design. Analysts spend less time cleaning data and more time framing questions. The people who lose ground are the ones whose value was concentrated in the automated part of the work — which is an argument for moving up the stack, not for panic.

How to move into one of these roles

Hiring managers for AI roles care much more about demonstrated systems than credentials. Ship something small and real: a retrieval pipeline over a dataset you know, an evaluation harness for a task you care about, a cost-optimized deployment. Write up what broke and how you measured it. That write-up outperforms most resumes.

Then apply where the roles actually are. Browse current openings on the GrowthAI job board and filter by the category you're targeting.

Keep reading

• How to Find Real AI Jobs (That Are Actually Hiring)
• How to Spot a Ghost Job Listing
• AI Job Titles Explained