Agentic AI — systems that don't just answer questions but take multi-step actions on their own, like booking, filing, querying databases, or triggering workflows — has moved from pilot projects to budget lines inside large enterprises over the past year. One of the clearest signs of that shift isn't a product announcement; it's a new set of job titles showing up on hiring platforms that didn't meaningfully exist two years ago.

AI Engineer is now the fastest-growing title, two years running

A Forbes roundup of emerging roles, compiling LinkedIn hiring data, lists "AI Engineer" as the fastest-growing job title in the US for a second consecutive year — distinct from the traditional "machine learning engineer" role in that it's oriented around integrating and orchestrating AI systems (including agents) into existing products, rather than training models from scratch. The same analysis attributes more than 1.3 million newly created roles to AI-driven demand, spanning well beyond pure research positions.

The rise of the Forward Deployed Engineer

Among the roles gaining the most attention is "Forward Deployed Engineer" — a title popularized by AI infrastructure companies for engineers who embed directly with a client to customize and deploy an AI agent against that client's specific systems and data, rather than building a generic product for broad release. It's a hybrid of software engineering, implementation consulting, and customer-facing work, and it's spreading beyond the companies that coined it as more enterprises buy agentic tools that need heavy customization to actually work inside a messy, real-world tech stack.

Supervision roles are catching up to building roles

Not all the new titles are about building agents — a growing share are about overseeing them. As agents are given more autonomy to take real actions (sending emails, modifying records, executing transactions), companies are creating roles focused on monitoring agent behavior, auditing decisions, and intervening when something goes wrong. This tracks with survey data showing a meaningful share of companies reporting that their deployed agents have already taken actions nobody explicitly authorized — a governance gap that's pushing "AI oversight" from a compliance afterthought into an actual job description.

A reality check on the hype

It's worth noting the adoption numbers behind these new roles vary sharply depending on who's measuring. A widely cited CrewAI survey found enterprises reporting near-universal plans to expand agentic AI use, while separate industry estimates suggest only around a quarter of organizations have actually gotten an agentic system into full production, with a large share of pilot projects expected to be shelved before they scale. New job titles are appearing ahead of mature, widespread deployment — which is normal for an emerging technology, but worth keeping in mind before treating every new title as proof of overnight transformation.

What this means if you're considering this path

• Forward Deployed Engineer-style roles reward people comfortable with ambiguity and client-facing work, not just strong coding skills — it's as much an implementation and communication job as an engineering one.
• AI oversight and governance roles are a realistic entry point for people with a risk, compliance, or QA background who want to move toward AI work without becoming a model developer.
• Because the field is still forming, job titles for essentially the same work vary a lot between companies right now — read the actual responsibilities in a listing rather than filtering by title alone.