47,000 job listings reveal the engineering roles that AI is creating

AI is not just changing what engineers do. It’s changing the job titles companies use to describe the work at all.

Andela’s research team reviewed 47,000 recent engineering postings from Fortune 500 companies and found more than 2,000 skills flowing into 23 emerging job titles. The pattern is familiar if you’ve watched tech long enough: older roles blur together under pressure, then a new label appears when the work gets specific enough to deserve one. DevOps did that. DevSecOps did that. Now AI is doing it again.

The headline finding is that these are not generic “AI” jobs. Among 1,832 postings aimed primarily at AI or ML engineers, 53% mixed skills from at least two established roles. The top new roles Andela identified include MLOps pipeline engineer, LLM application engineer, FinOps reliability engineer, docs-as-code engineer, and product frontend engineer. Each one combines pieces of older jobs, but for a narrow operational need: model deployment, model evaluation, cloud cost and reliability, documentation that behaves more like code, or frontend work tied closely to product decisions.

Cory Hymel, Andela’s head of research, argues that the data cuts against the idea that AI will make engineers more generalist. Quite the opposite. The jobs are getting more specialized, not less, and some of the fuzzy middle is being pushed toward automation. The skills that remain closest to the core of a role matter more when AI can take over some of the cross-role work.

That has practical consequences for hiring. Hymel says enterprise job descriptions are often too generic, especially when HR starts the process far from the actual work. He points out that companies can end up with dozens of front-end listings that all ask for different things, then wonder why the candidate pool is messy. His fix is blunt: describe the outcome first, separate required skills from preferences, and stop letting AI write the most human part of the hiring process.

For engineers, the advice is equally specific. Product-minded people should look at product frontend roles. Docs writers may want to move toward docs-as-code work. And the broader message is that AI is not flattening tech jobs into one big mushy title. It’s carving them into cleaner pieces, whether companies are ready for that or not.