Designs scalable QA automation and AI evaluation frameworks across web, APIs, data pipelines, and LLM-powered features. Establishes risk-based testing, quality gates, CI/CD integration, observability, defect triage, and non-functional security testing. Partners with engineering, data science, product, and SRE teams to enable shift-left quality ownership. Evaluates AI systems for hallucinations, prompt injection, drift, retrieval failures, and agentic workflow risks using golden datasets, semantic scoring, and LLM-as-a-judge approaches.