Kai Greshake

Kai Greshake is a researcher and consultant focused on the security of large language models and AI systems. Formerly a penetration tester, he now works at NVIDIA to make AI deployments safer and more resilient. His research on prompt injection attacks, including indirect prompt injection, has been widely cited. Kai explores how complex systems break—and how to rebuild them more securely. He shares his insights through research papers, open-source projects, and writing on AI security.