AI safety · Evaluation
ServalGuard® AI Evaluation
Designed a four-pillar evaluation method and ran 400+ adversarial tests.
Built independently · Houston, TX
All artwork shown here belongs to me. I designed, built and delivered these products, and created their visuals too.
AI safety · Evaluation
Designed a four-pillar evaluation method and ran 400+ adversarial tests.
AI safety evaluation and child-protection tools · 2023 to present
Built AI safety products, detection bots and published field research.
AI agents · Community support
Developed free AI agents for community safety and practical support.
Interactive safety game
Created Safi’s world and built a playable safety mystery.






Research · Tooling
Built attack suites and mapped safety risks across chatbot conversations.
Nonprofit · Research
Founded a nonprofit research center focused on digital harm prevention.
Developer · Apps
Designed cooking and neurodivergent-user app concepts.