Mo Ragiv R.

Senior QA Automation Engineer / SDET with 5 years of experience, now focused on testing AI solutions, evaluating AI models, and building AI-powered testing systems. If you're shipping an AI product or LLM feature, testing it isn't the same as testing regular software. Outputs vary, "correct" is a judgment call, and quality has to be measured, not just asserted. That's the gap I fill. I evaluate model responses for correctness, efficiency, and safety, and I build the automation and frameworks that keep AI features reliable as they change. What I can do for you: 🔹 Evaluate AI models and LLM outputs — rubric-based scoring, side-by-side model comparison with clear written justification, RLHF annotation, and flagging when a model drifts from the task 🔹 Test AI-powered features — prompt and output validation, regression checks for non-deterministic responses, and catching where a model quietly does the wrong thing 🔹 Build AI testing solutions — I built an AI-driven test-generation system on the OpenAI API and a self-healing test framework that repairs its own locators, cutting maintenance 🔹 Own end-to-end QA automation — web, mobile, and API, wired into CI/CD so tests run on every commit 🔹 Build stable, maintainable frameworks using proven patterns (Page Object, Factory, Strategy) that hold up as the suite grows A production-data replication tool I built cut test setup time by ~80%, so I care about results, not just coverage. Domain experience: AR/VR devices in social media & entertainment, fintech (mortgage loan processing) and retail. I work smoothly with international teams across time zones and communicate clearly in English. Tools I work with Languages: Python, JavaScript, SQL Automation: Selenium, Playwright, Appium, Pytest, Behave, Cucumber, REST API testing CI/CD: Jenkins, GitHub Actions, Azure Pipelines Test & project management: TestRail, Jira, Azure DevOps, ClickUp AI & AI-assisted QA: OpenAI API, Claude Code, Cursor, LLM evaluation, RLHF workflows Need someone who can test your AI, measure how good your model really is, or take AI-powered QA off your plate end to end? Send me a message and let's talk through what you need.