Agentic QA (Intermediate QA)

Salary: $75,000 – $90,000 per year

About the role

This is the second quality seat in the engineering team. Our engineers build through agentic/ai practices, and when code gets easier to write and no cheaper to trust, the constraint moves from building to verifying. This is an intermediate role where you direct testing agents to carry the volume and bring the judgement they do not have. You work across what we ship: the core platform and its integrations, the web interface, and the mobile work as it comes online, as one of two quality people here.

Key responsibilities

• Take a change, work out what it puts at risk, and cover it properly through independent verification

• Direct agents to draft and maintain tests, widen coverage, and take the first pass at a failure, wired into the browser and the build

• Set what testing agents may touch, where they stop, and what has to come back to a person

• Read what agents produce, understand it, and ensure you never ship a check you could not explain

• Add to and keep regression suites healthy: fast, stable and trusted, with flaky tests dealt with rather than tolerated

• When something breaks, get to the bottom of it, working back through the trace and the logs to determine whether it is the product, the test or the environment

• Turn a vague report into a failing test an engineer can work from

• Leave behind evidence of what was tested, by what, and why the result can be trusted

• Work on testing AI features that do not give the same answer twice, defining a pass by the right shape, grounded in the vehicle or the lead in front of it

About you

• Around three years of hands on testing and test automation, including time before agentic tools existed

• You already work with agentic tools, every day, and have wired them into real work, driving a browser and a build through something like the Model Context Protocol

• Solid test automation: end to end on web with tools such as Cypress or Playwright, some API testing, and cross browser and device coverage

• You can write a suite, keep one healthy, and set out what a test should do in a form a person and an agent can both read

• Strong on failure analysis: traces, screenshots and logs, and the judgement to tell a real defect from a flaky test, a broken environment or an intended change

• Genuinely good at exploratory testing and finding the edge case nobody wrote down

• Comfortable against a .NET backend and its API surface, including the older integrations that are not tidy

• Familiar with continuous integration, release pipelines and git based pull request workflows, including reviewing one a machine wrote

• You can read a change and reason about what it touches, rather than waiting to be told what to test

• Clear written English and a feel for the wider quality picture: safe handling of credentials, dependencies and customer data

• The right to work in New Zealand

About us

Autoplay is the lead management platform that car dealerships, dealer groups and vehicle manufacturers across New Zealand and Australia run their sales floors on. Every enquiry is tracked from first contact through to vehicle handover, and we work with dealers of every size, from single sites through to groups and manufacturer networks, on both sides of the Tasman. We are part of Vela APX, a group of automotive technology businesses operating across Australasia. Engineering sits in Auckland and is deliberately small, so what we protect is the quality of what we ship rather than the size of the team shipping it.

Benefits

• A team small enough that what you change on a Tuesday is how we work on a Wednesday

• AI tools in the hands of everyone here, used to reach further rather than to finish earlier

• Auckland based, in person with the team, with the flexibility you would expect

• Vela APX group benefits: bonus and referral schemes, share purchase plan and an employee assistance programme