Skip to content
Trellis

Applied AI Engineer

Team
Engineering
Location
Oakland, CA or remote in the United States
Type
Full time
Compensation
$180,000 to $230,000 plus equity

You will own the model layer: routing, prompts, evaluation and the verification that keeps our output honest. The hard part here is not generating text, it is proving that a claim about a customer's evidence is grounded in the customer's evidence.

What you will own

  • Own routing across providers and the evaluation suite behind it
  • Build the grounding checks: quote verification, evidence attribution, refusal to guess
  • Cut cost per readiness run without moving quality
  • Keep the routing table honest and published

Who this suits

  • You have shipped a language model feature that customers rely on
  • You have opinions about evaluation that came from being burned
  • You are comfortable saying the model is wrong and building the check that catches it

If you read the list and match most of it, apply. We have hired two people who did not match a bullet each and both were right to send the form anyway.

Apply for this role

Goes straight to a founder. We reply either way, usually within a week.

Or email careers@trelliscore.co with the role in the subject.