Agentic Automation Engineer (Remote)

**About Keeyu** Keeyu is an AI\-powered e\-commerce operations platform. We use agentic workflows, LLM\-based tooling, and event\-sourced architecture to run shipping, fraud, order management, and customer operations for brands like Decjuba, Rebel Sport NZ, and Budgy Smuggler. We're a small, fast\-moving team, backed by Rampersand and Startmate, and we have strong opinions about staying lean — we automate aggressively rather than throwing headcount at problems. **About the role** This is a role for someone who looks at any manual or repeating task and instinctively asks, "why is a human still doing this?" You'll own automation across three areas of the business — testing/QA, customer onboarding, and production reliability — building agentic systems and tooling that make manual work disappear. You won't be handed a backlog of tickets. You'll go find the bottlenecks, the flaky steps, the issues that page someone at 2am, and the onboarding tasks that still need a human in the loop — and you'll automate them. *\-\- This is not a QA / Tester role, its much more fun than that \-\-* **Responsibilities** * **QA \& Testing automation** Build and own automated testing across the whole stack — frontend portals, backend APIs, and our AI agents and the tools they use. This goes beyond traditional test suites: you'll figure out how to reliably test non\-deterministic agentic behaviour, validate tool outputs, and catch regressions across the platform before they reach customers. * **Customer onboarding automation** Most of our onboarding is already automated, but the gaps are where the pain lives. You'll map the full onboarding journey (new customers vs. existing/migrating customers), find the manual steps and bottlenecks that still slow things down, and automate them away so onboarding scales without scaling effort. * **Production reliability \& observability** Hunt down repeating production issues and kill them at the root. You'll build proactive observability, detection, and self\-healing automation so recurring deployment and environment issues get resolved automatically — ideally before anyone notices them. **Requirements** * Genuinely curious — you can't leave a manual process alone once you've spotted it * Very comfortable building with AI agents, agentic workflows, and modern automation tooling (core to this role, not a nice\-to\-have) * A programmer or automation tester by background — you write real code and build robust, maintainable automation rather than brittle scripts * Excellent communication skills * Able to work Australian business hours with a minimum of 4–5 hours of daily overlap with our Sydney team **Nice to have** * Experience testing or building LLM\-based systems and evaluating non\-deterministic outputs * Familiarity with event\-driven / event\-sourced architecture * Experience with CI/CD pipelines, container orchestration, and cloud\-native infrastructure (we're Azure\-native — AKS, APIM, and friends) * E\-commerce or operations platform experience * A track record of building observability and incident\-prevention tooling **Why join** * Real ownership — you'll define automation strategy across the business, not just execute a backlog * A team that genuinely values lean, pragmatic engineering over busywork * Fully remote, built around shipping fast and automating relentlessly * Work at the frontier of agentic systems on a platform already running in production for real brands