Test Engineer - AI Software QA

About the role

Rehabitaire builds AI-powered physical therapy software that connects healthcare professionals and patients with real-time feedback, personalized exercise programs and progress tracking. This Test Engineer position sits inside the AI team, working daily with AI engineers, data annotators and the product owner. The role owns the release gate, continuous integration, test data, live testing and load/stress testing, and is based on site in Aleppo.

Key responsibilities

• Own the release gate: turn acceptance rules into code and run automated accuracy tests on held-out, product-owner and live-session test sets, giving a go/no-go recommendation.
• Build and run CI pipelines so unit tests run automatically, including integration tests, API tests, browser end-to-end tests and a smoke test after every deploy.
• Keep golden test sets frozen and versioned with the data annotator, automate annotation checks, and turn field bugs into regression clips while respecting privacy rules.
• Write protocols for live test rounds (camera angle, distance, lighting, mirrored front cameras, devices) and turn findings into reproducible tickets.
• Run load and stress tests on servers, pods and GPUs, tracking latency, frame rate, GPU memory and errors, and rerun them before releases that touch them.

Required profile

• 3+ years in test automation, or as a software engineer who owned the tests of a product.
• Strong Python, including pytest, fixtures and mocking, testing a Django / REST backend.
• Experience building and running CI pipelines (GitLab CI, GitHub Actions or similar) with Docker.
• Comfort with evaluation metrics such as precision, recall, F1, IoU and confusion matrices, and the judgement to separate real changes from noise.
• Browser end-to-end tests (Playwright or Cypress) for a React app.
• Clear bug reports, reproducible steps and short test reports.

Nice to have

• Testing of machine learning or computer vision systems, especially video or pose estimation.
• GPU workloads, ONNX Runtime or PyTorch inference.
• MongoDB, PostgreSQL, Redis, Celery and cloud infrastructure and deployment.
• Real-time video (WebRTC, LiveKit).
• Load and stress testing tools such as k6 or Locust.