@kingroryg

work

I help raise the security bar for Amazon Sagemaker at AWS. I work mostly in network protocols, applied cryptography, and ML security.

research

Traceable Content Generation via Activation Steering [arXiv 2025]

Rethinking Evals: Behavioral Attraction Basins in Language Models [arXiv 2025]

MAPLE: Specialized Modular Neural Architectures for Multi-Agent Reinforcement Learning [PDF]

ACSE-Eval: Can LLMs threat-model real-world cloud infrastructure? [ESORICS Workshop 2025]

Representation Engineering for Large-Language Models: Survey and Research Challenges [ACM Survey 2025]

security

tba

other work

blog posts