AI-Assisted Annotation and Harm Measurement Specialist (King’s College London)
Text
Led the design and execution of AI-assisted workflows for measuring online harms using structured annotation pipelines. Integrated human review, LLM-assisted scoring, and adjudication into evaluation frameworks targeting extremist content. Developed scalable and auditable systems to support detection and safety metric generation in multimodal online environments.
• Built structured annotation pipelines enabling multi-annotator adjudication.
• Developed evaluation frameworks combining LLM scoring and human judgment.
• Conducted actor- and network-level harm annotation and mapping.
• Delivered operational harm measurement pipelines for real-world enforcement.
Led the design and execution of AI-assisted workflows for measuring online harms using structured annotation pipelines. Integrated human review, LLM-assisted scoring, and adjudication into evaluation frameworks targeting extremist content. Developed scalable and auditable systems to support detection and safety metric generation in multimodal online environments. • Built structured annotation pipelines enabling multi-annotator adjudication. • Developed evaluation frameworks combining LLM scoring and human judgment. • Conducted actor- and network-level harm annotation and mapping. • Delivered operational harm measurement pipelines for real-world enforcement.
2026 - Present