Skip to content
Over Unity

AI engineering

Builds the systems that put models into production and keeps them there.

Why this one is hard to judge

A demo proves nothing. What matters is what they did when it failed under real load.

What to ask for

  • Ask for a system they built that is still running in production today.
  • Ask what happens in their system when a model call fails or times out.
  • Ask for the cost per request of something they shipped, and how they reduced it.
  • Ask them to describe a bug that only showed up after real users touched it.

The mistake most hirers make

Hirers judge candidates on how well they can explain a model, not on how they handle the mundane parts around it. The interesting work is in retries, logging and fallbacks, not in the prompt. Someone who only demos a notebook has not built anything a business can rely on.

What good looks like after 90 days

A working feature in production, not a demo, with basic monitoring around it. At least one incident handled calmly, with a clear account of what went wrong. Other engineers can read their code without asking them what it does.

How we assess it

Against a rubric that is published in full, on evidence the practitioner supplies and a reviewer checks. Where something has not been verified, the profile says so.

Find someoneRead the rubric

Over Unity makes introductions between hirers and independent specialists. It is not a party to any engagement, does not hold or transfer payments, and does not determine employment status. Specialists are never charged a fee.