Engines: the intelligence every Hub is built on.
An Engine is reusable intelligence — reasoning, search, vision, or automation — built once and shared by every Hub that needs it, instead of rebuilt separately inside each one.
A capability, not a product
An Engine is a specific kind of intelligence — reasoning about a document, searching across sources, reading an image, or automating a sequence of steps. On its own, an Engine isn't something a person opens and uses directly; a Hub composes one or more Engines into an actual workflow, the same way Resume Optimizer composes reasoning and matching to turn a resume and a job description into a tailored rewrite.
Built once, not six times
Without a shared Engine layer, every Hub would need its own version of the same reasoning and matching logic — six implementations of roughly the same idea, improving at six different rates and drifting apart in behavior over time. Engines exist so that work happens once, in one place, at a quality every Hub can depend on.
Below Hubs, informed by Labs.
Hubs call into Engines to do their actual work. Labs is where new Engine capabilities get tested before they're trusted enough to power a live Hub — the two are connected, but a Labs experiment isn't an Engine until it's proven out.
Two live products, both in Career Hub.
Resume Optimizer
ATS scoring, keyword matching, and job matching are all reasoning-Engine work, applied to one document at a time.
Interview Coach
Reads a candidate's answer and generates specific, direct feedback — the same reasoning Engine, a different prompt and workflow around it.
Quality that compounds instead of fragments.
Consistent behavior, not six different implementations
A scoring model built once and reused is easier to trust, measure, and improve than the same logic rebuilt slightly differently in every Hub.
Improvements reach every Hub at once
When an Engine gets better, every Hub that depends on it benefits immediately, without a separate update for each one.
Labs → Engines → Hubs
Research that starts in Labs, once it's proven reliable enough, becomes part of an Engine. Engines sit in the middle of the platform: below the Hubs that depend on them, above the raw research that produced them. A Hub never talks to a model directly — it talks to an Engine, and the Engine is what actually does the work.
See an Engine at work
Resume Optimizer's ATS scoring and job matching are Engine capabilities you can try today.