Trusted Tester & Red Team Contributor
Don T DiscloseText
Contributed structured adversarial inputs and safety evaluation reports to trusted tester programs for frontier LLMs and multimodal AI models. Identified safety gaps in language, code, image, video, and speech systems to inform responsible scaling. Assessed generative systems pre-deployment and provided recommendations for policy compliance and robustness.
• Tested LLMs, code generation, and multimodal models for safety and fairness.
• Delivered structured feedback across text, image, speech, and video modalities.
• Identified vulnerabilities in diffusion and generative models before launch.
• Supported rollout of safety mitigations and harm reduction best practices.
Contributed structured adversarial inputs and safety evaluation reports to trusted tester programs for frontier LLMs and multimodal AI models. Identified safety gaps in language, code, image, video, and speech systems to inform responsible scaling. Assessed generative systems pre-deployment and provided recommendations for policy compliance and robustness. • Tested LLMs, code generation, and multimodal models for safety and fairness. • Delivered structured feedback across text, image, speech, and video modalities. • Identified vulnerabilities in diffusion and generative models before launch. • Supported rollout of safety mitigations and harm reduction best practices.
2024 - Present