Solvea
Careers
AIQ bar

What strong AI-native evidence looks like

We are not filtering for resume polish. We are looking for proof that you can use AI to think, build, judge, and ship.
Onsite · San Jose Evidence over credentials Taste over prompt fluency

The bar

Strong signal

A-class candidates show the work, the loop, and the judgment.

  • Working artifact, shipped workflow, repo, demo, review, or real customer-facing output.
  • Before/after evidence showing weak first output and the stronger final version.
  • Clear tool loop: Claude Code, Codex, agents, memory, evals, scripts, or feedback loops.
  • Judgment: what you rejected, corrected, measured, and would rebuild next.

Weak signal

These usually fail the AIQ screen, even when the resume looks good.

  • Generic claims like "I use AI every day" without artifacts or outcomes.
  • Only screenshots of prompts, with no quality bar or before/after reasoning.
  • Credential-heavy applications that do not show how AI changed the work.
  • Polished writing with no evidence of taste, correction, shipping, or learning speed.

Submission examples

Builder example

Build a small GTM agent that researches a market, identifies target customers, drafts outreach, and records where the agent was wrong.

  • Submit the demo, repo, or screen recording.
  • Include the first bad output and the final improved workflow.
  • Explain what you measured and what you would automate next.

Apply as AI-Native Builder →

Reviewer example

Take weak AI output from a campaign, landing page, support reply, or VOC summary. Diagnose it, improve it, and make the taste bar explicit.

  • Show the original output, review notes, and final version.
  • State what was misleading, generic, off-brand, or unshippable.
  • Explain the rubric you used to judge quality.

Apply as AI-Native Reviewer →

Servicer example

Turn an AI workflow into real customer trust: a reply, escalation path, follow-up sequence, operating playbook, or field-ready handoff.

  • Show the customer context and the AI-assisted workflow.
  • Explain how you prevented wrong or low-trust responses.
  • Include the final message, process, or handoff that could ship.

Apply as Servicer →

Fast path: start with the AIQ Challenge. If the evidence and AI interview show A-class signal, you can book Hunter directly at book.paircode.ai/hunter. Interviews are in San Jose.