• Led GenAI, RLHF, and LLM safety evaluations and evaluation workflows across Google AI ecosystems, including Gemini-aligned testing and high-risk model assessment workflows.
• Developed adversarial prompts, red teaming workflows, and AI safety evaluations to identify harmful outputs, model vulnerabilities, and policy violations.
• Conducted multi-turn LLM evaluations to assess model alignment, child safety compliance, hallucination risk, and unsafe GenAI behavior.
• Evaluated AI outputs against RAI, RLHF, and T&S standards to support launch readiness, mitigation strategies, and go/no-go decisions.
• Partnered with Eng, Product, Policy, and Ops teams to improve AI safety guardrails and model risk mitigation.
• Tracked risk trends, escalations, and evaluation metrics using operational analytics and reporting workflows.
• Directed vendor GenAI safety evaluations and RLHF review standards for child safety compliance.
