Director, Research - AI Evals

Own how Figma measures the quality of its AI-powered experiences. Define what “good” means, build evaluation frameworks, rubrics, golden datasets, and quality bars that combine human judgment with automated/model-based methods (including LLM-as-judge). Partner with engineering to create repeatable evaluation pipelines and regression testing, and deliver dashboards/readouts that support go/no-go decisions while leading a small team.