Product Engineer

Arbitr Group Ltd; Upper Harbour; On-site; NZD 120,000 - 180,000; Full time

Job summary

Arbitr Group Ltd in Auckland's Upper Harbour seeks a Product Engineer to orchestrate AI-powered delivery within autonomous Pods. You’ll translate briefs into production-ready software by directing AI tools, writing tests, and validating outcomes at high velocity.

You’ll own end-to-end work, master prompts, verify AI output, and push for edge-case resilience. This role demands deep engineering fundamentals and a penchant for rapid, responsible experimentation with AI.

Qualifications

• Recent experience shipping production software with AI tools.
• Write precise prompts and iterate quickly when outputs miss.
• Evaluate AI-generated code with rigorous testing and validation.

Responsibilities

• Go AI-first: use AI tools as the primary mechanism for feature delivery.
• Own the full stack end-to-end from brief to deployment inside your pod.
• Master prompts and maintain a log for significant features and outcomes.
• Verify AI outputs with stress-tested test suites and edge-case checks.
• Identify hallucinations, logic gaps and fix them before production.
• Stay updated on AI techniques and bring them to the team.

Skills

AI tooling

Prompt engineering

End-to-end delivery

Test automation

Systems thinking

Tools

Claude

Cursor

Pi AI

Job description

We're not hiring coders. We're hiring orchestrators of outcomes.

arbitr runs on Pods — small, autonomous teams built for maximum speed. As a Product Engineer, you're the primary delivery engine inside one of them. You take a product brief — and use advanced AI orchestration to design, build, test, and ship at a pace traditional engineering simply can't touch.

This is a High-Velocity execution role, and we mean that literally. We don't need you to type every line of code — we need you to direct AI to produce it, verify it with surgical precision, and ship work that's genuinely production-ready. In our world, velocity without verification is a failure. We expect both, every time.

We're only interested in the best of the best. If that's an uncomfortable bar, this isn't your role. If it's the bar you've been looking for, keep reading.

What you'll do

• Go AI-first. Use Claude, Cursor, and Pi AI agent as your primary mechanism for building features — not as occasional assistants you dip into.
• Own the full stack. Translate product briefs into working software end-to-end, from initial design through to deployment, inside your pod's domain.
• Master your prompts. Keep a prompt log for significant features — how you directed the AI, where it broke, and how you steered it back on track.
• Verify like it's your name on it — because it is. Write and critically stress-test AI-generated test suites so they challenge the logic, not just wave the happy path through.
• Hunt the failures. Actively go looking for hallucinations, logic gaps, and edge cases in AI-generated output before any of it gets near production.
• Stay on the frontier. Show up to pod rituals, chase the newest AI techniques as they land, and bring them back to raise the whole team's floor.

What you'll bring

You have real, recent experience shipping production software with AI tools — not side projects, not "Hello World." You write precise prompts and iterate fast when the output misses. You measure yourself by what ships and the value it creates, not the hours you spent typing.

You treat AI as a capable, fallible collaborator — never gospel. You can spot shallow error handling, missing edge cases, and overconfident logic from a mile away, and you know exactly when to trust the output versus when to roll up your sleeves and dig into the architecture yourself.

You've got real engineering fundamentals underneath it all — enough to know instantly whether AI-generated code is structurally sound or a house of cards. You've let go of the manual habits AI now does better, without losing the judgement those habits were built to protect. And you're comfortable starting from a rough brief and iterating toward an answer, rather than waiting for someone to hand you a perfect spec.

Why this role

You'll operate inside a Pod built for autonomy and speed, spend far less time on boilerplate, and far more time on the high-leverage problems — using the most advanced AI tools on the planet to build the future of Straker and arbitr.

This isn't a startup pitching a dream. It's a real, global, ASX-listed company backing a new era with everything it's got — and it's looking for the engineers sharp enough to build it at AI speed.