At Scale AI, I audit generative AI training datasets and annotator outputs for accuracy, reasoning, grammar, tone, and rubric compliance. I maintain gold-standard examples, honeypots, and edge-case FAQs, and uphold a 98%+ audit accuracy rating.
At Outlier.ai, I conducted RLHF and pairwise evaluations using Helpfulness, Harmlessness, and Honesty rubrics. I reviewed multi-turn reasoning and factual claims to identify hallucinations and provide alignment critique.
As an independent AI consultant, I audited multi-turn dialogs, synthetic voice prosody, and computer vision datasets across several platforms. At HumanSignal, I validated document layouts, slide presentations, and text-image alignment for training benchmarks.
Across these roles, I’ve supported contributors through onboarding, rubric guidance, and written feedback. I’ve also created style guides and QA documentation and used Discord and Google Sheets trackers to support remote project operations.

