Senior Automation Testing Engineer

The Role

Test automation across the stack:

• Backend: API and contract testing, service-level and integration coverage, data setup that doesn't rot, and test design that survives a schema change.
• Frontend: web E2E and component-level coverage with Playwright, visual and RTL regression, and suites fast enough to gate a merge rather than a nightly.
• Mobile: native and cross-platform coverage with Appium or Maestro, device-farm strategy, offline and sync behavior, and the payment-peripheral paths that only break on real hardware.
• The connective tissue: shared fixtures, environment and test-data management, parallelization, and CI pipelines where a red build means something.
• Performance and load testing experience.

The AI layer on top of it:

• Test generation from specs, code, and production traffic — with the maintenance story solved, not just the first draft.
• Failure triage that classifies a red build before a human opens it: real bug, flake, environment, or test rot.
• Self-healing locators and suite health tooling — flake detection, quarantine, coverage-gap analysis.
• Evaluation infrastructure for AI features across our products: datasets, scoring, and regression detection when a prompt or model changes.
• Evaluation for our market specifically — Arabic and English behavior, RTL interfaces, and region-specific POS, tax, and payment rules. Correctness here is rarely a string match.

Agentic AI and orchestration:

• Agentic AI that does real work in our pipelines: reads a diff, runs the relevant suite, reproduces a failure, proposes a fix, opens the PR.
• Agents that own a quality workflow end to end — exploratory testing against a running build, coverage-gap hunting, release-risk assessment — and know when to escalate to a human.
• Orchestration that holds up under load — multi-step planning, tool use, retries, state and memory across steps, sandboxed execution, multi-agent handoffs, and clean boundaries between agentic and deterministic steps.
• Integration with the stack we already have (CI, Jira, observability, MCP-style tool interfaces) rather than a parallel system beside it.
• The judgment to know when a plain pipeline beats an agent, and to say so.

The technical ground:

You should be current on how this work is actually done today, and able to argue about it rather than recite it:

• Test automation: framework design and layering, the test pyramid and where it stops being useful, flake economics, parallel execution, mobile and cross-browser realities, CI/CD gating, Playwright, Appium, Maestro.
• Agentic AI: orchestration and tool use, multi-step planning, memory and state, sandboxed execution, multi-agent patterns, MCP and similar tool-integration standards, and the cost of each.
• Context engineering: retrieval strategy, chunking, reranking, caching, and managing long-context behavior — including where it degrades.
• Evaluation: offline and online evals, LLM-as-judge and its failure modes, human-in-the-loop review, statistical significance on small samples, regression gates in CI.
• Reliability: structured output, guardrails, fallback and retry design, and handling non-determinism in systems that must not flap.
• Operations: tracing and observability for LLM systems, prompt and version management, latency and cost budgeting, model routing, and when fine-tuning or distillation beats a better prompt.

Requirements

• An engineer who ships production software, with recent hands-on work on LLM-backed systems that real users depend on. Strong Python; comfortable in at least one of .NET, Java, or TypeScript. Tested, maintained code — not notebooks.
• Real automation depth across more than one surface. You've owned a suite that gates releases on backend and on a UI — web or mobile — and you can explain how you kept it green without deleting the hard tests.
• Real experience building evaluation systems. You can explain how you knew your system was getting better, with numbers.
• Practical depth with the modern LLM toolkit — prompting, structured output, tool use, retrieval, agentic AI orchestration — and a clear sense of the trade-offs.
• Credible testing fundamentals. You don't need a QA title, but test design, automation frameworks, and CI/CD shouldn't be new to you.
• A bias toward adoption. You measure your work by what other engineers use, not by what you demoed.

Benefits

• Highly competitive compensation packages, including bonuses and the potential for shares.
• Regular training and an annual learning stipend to tackle new challenges and grow your career in a hyper-growth environment.
• Join a talented team of over 30 nationalities working in 14 countries, and gain valuable experience in an exciting industry.
• Autonomy, mentoring, and challenging goals that create incredible opportunities for both you and the company.