Case studies
Frontier AI training & evaluation
High-signal training for OpenAI, Anthropic (cybersecurity), and Google Gemini.
Overview
Delivered high-signal training and evaluation work across frontier model programs, with rubrics and tooling for consistent expert review.
Outcomes
Improved model quality on expert technical and safety tasks.
Summary
High-signal training for OpenAI, Anthropic (cybersecurity), and Google Gemini.