Skip to content
SCALECRAFT Advisors
← All roles

Agent testing & evals

QA / Evaluation Engineer

Prove the agent works before it goes anywhere near a real client — with numbers, not opinions.

Freelance, project-based · billed per sprint or hourly

Objectives

  • ·Design a test and evaluation harness covering the agent's core and edge-case scenarios
  • ·Measure task success rate, hallucination rate, latency, and cost per task
  • ·Flag failure patterns clearly enough for the build developer to act on

Key deliverables

  • ·A written test plan mapped to the agent specification
  • ·An evaluation report with pass/fail metrics
  • ·A signed-off go/no-go recommendation before deployment

Experience needed

  • ·Experience testing LLM-based or agentic systems specifically, not just traditional QA
  • ·Comfortable writing test scripts, Python preferred
  • ·Familiarity with eval frameworks, or willing to build a lightweight one from scratch

What to send

  • ·Send your CV.
  • ·Include 2–3 relevant past projects — the specific role you played, the tools you used, and the outcome. Not just a job title.
  • ·Give us one reference contact from a relevant project: name, relationship, and email or phone.
  • ·If you can cover more than one role, please state this clearly in your cover letter.

Fees are project-based and discussed openly once we see a fit — tell us your usual rate in the message if you would rather lead with it.

Apply

For QA / Evaluation Engineer. Everything you send goes straight to the founder.

2–3 relevant projects — your specific role, the tools, the outcome. Plus one reference contact we can reach.

CV and any work samples. PDF, Word, PNG or JPG — up to 10 files, 10 MB each.