Which AI role fits you — and how ready are you today?

In Java-shop hiring, "backend engineer" is one ladder with one loop. AI hiring in 2026 is not that: five different roles share the word "AI" in the title and are tested, and paid, completely differently — a system-design round for one is a customer-discovery role-play for another. Tap a role below for what it's actually hired for, its loop, and the site path that prepares for it. Then scroll down: the readiness bars are computed live from your own progress on this site, not a self-report.

5 roles sourced 2026-09-27 reads The Path's progress ~6 min

Five roles, five loops

Every card below cites the research on what top companies want (2026). Tap a role; the panel below fills in with what it's hired for, how it's tested, and where on this site to go next.

Pick a role

👉 Tap a role card; predict how its loop differs from a plain backend loop before you check the readout.

AI Engineer

Ships LLM/RAG/agent features to production.

Applied AI / FDE / Solutions Architect

Builds AI systems directly with a customer.

AI Architect

Designs the enterprise system around the model.

ML Engineer

Owns training and eval pipelines, not just API glue.

Research Engineer

Implements and extends papers at speed.

Tap a role above.

Takeaways. All five roles say "AI" and none share a loop. The fastest way to waste interview prep is practicing the wrong one — check the readout above against the job posting's actual day-to-day before picking which milestones below to chase.

Your readiness

Computed from this browser's own record of what you've done — the same store The Path uses, plus the Architect and Threat-Modeling tracks' local progress. Nothing is uploaded; open this page in a different browser and it starts from zero.

Loading your progress…