
AI Model Evaluation: Fact-Grounded Tests for Frontier Models
Designed question-answer evaluation tasks that expose failure modes in frontier AI models, each anchored to a single verifiable fact and checked through a multi-stage peer review.
AI evaluation, language data, and workflow automation

Designed question-answer evaluation tasks that expose failure modes in frontier AI models, each anchored to a single verifiable fact and checked through a multi-stage peer review.

Annotated and validated speech datasets for machine learning models, including prompt engineering, QA workflows, and model calibration.

A version-controlled LaTeX pipeline that assembles tailored cover letters and application packets from reusable modules, extended with Claude Code agents and skills that check documents for ATS problems and keep the repository clean.

Designed, built, and deployed a business website with Hugo, GitHub, and Netlify, including a custom audio-library component for browsing a large catalogue of voice and music samples on mobile and desktop.

Designed, built, and analyzed a large-scale sociolinguistic dataset, including annotation frameworks, QA workflows, and statistical modeling.