Hire AI engineers on what they've shipped.

Hiring AI engineers is hard because the usual filters don't work in a field this young. ChatGPT came out on November 30, 2022. A job post that asks for ten years of LLM experience screens out everyone, including the people who can do the job.

So recruiters fall back on what they can check fast: degrees, years, big-company logos. None of that tells you whether someone can put an AI automation in production. Pedigree is a proxy. In AI, it's a bad one.

We recruit two kinds of people: AI engineers, who build the systems, and AI operators, who find where AI should be doing the work and ship the automations. Most companies need both.

We judge AI candidates on what they've shipped, not on résumés or pedigree. A résumé tells you where someone sat. Shipped work tells you what they can do.

• Shipped work over résumés. A product, an automation, or a repo that real people use says more than a degree or a job title.
• Recent work over years. Nobody has a decade of experience shipping LLM products, so counting years filters out the wrong people. What someone built this year matters more.
• Anywhere over nearby. We recruit globally, so the pool is everyone who has shipped, not everyone who lives near your office.

Remote hiring lets you choose from the whole world instead of whoever lives within commuting distance of your office. In a field this young, the person who has already shipped what you need is rarely down the street.

It only works if you know how to run remote teams. We've run them since 2003:

The full story is in Alex's profile, Tania's profile, and the origin story on the home page.

Hire when you already know what you want built. When you don't, start with a 90-Day Placement, or with an AI roadmap if you only need the map. A new hire with no map starts by learning where the hours go. An operator on a placement is there to do exactly that: days 1–30 are for mapping.