Micah Bickham

Evaluated multimodal model outputs using strict rubrics and objective evaluation criteria; assessed image and audio fidelity for accuracy, safety, and guideline compliance; documented decisions for auditability on annotation platforms; labeled sentiment, quality, and relevance for text/search/ads/assistant responses; consistently exceeded QA targets while maintaining strong throughput; resolved edge cases by refining rubrics and taxonomies, improving inter-rater agreement and reproducibility; detected labeling errors, proposed clear fixes, and wrote rationale notes that aligned with objective evaluation criteria; conducted targeted research (incl. AI tools) to support well-justified ratings and contributions to dataset curation/quality.