Independent AI Training & Annotation Specialist (Remote, Contract)
Evaluated and ranked AI-generated responses across domains such as creative writing, factual summarization, and logical reasoning to improve foundational LLM outputs. Designed multi-turn adversarial prompts to stress-test model boundaries, detect hallucinations, and enhance response safety and helpfulness. Performed quality assurance to support consistent high performance on annotated tasks for technology clients. • Response ranking and evaluation of AI outputs • Red-teaming via adversarial prompt design • Hallucination identification and safety/helpfulness improvement • QA-driven review against evolving rubrics and guidelines