Summary
Erick Galinkin is an AI security researcher with seven years of hands-on experience uncovering and mitigating vulnerabilities in large language models and cloud systems, currently building security guardrails for LLMs at NVIDIA. He previously founded and led AI research at Rapid7, producing public threat research, zero-day exploits and detections, and advising industry and government on AI policy. His background spans offensive and defensive roles—from reverse engineering malware and IDS content at Cisco to cloud and devops security at Optiv and Netskope—giving him a rare perspective that blends exploitation, detection engineering, and operationalization. Erick pairs deep academic training in computational math, information assurance, and bioinformatics with practical instruction experience (Udacity, UC Berkeley, Harvard) and service in the U.S. Air Force, and he speaks Chinese, an asset he has used in threat intelligence work. Not obvious from his job titles: he frames his research as “doing bad things to good models,” reflecting a pragmatic red-team ethos that focuses on making AI systems safer by thinking like an attacker.
7 years of coding experience
9 years of employment as a software developer
Doctor of Philosophy - PhD, Computer Science, Doctor of Philosophy - PhD, Computer Science at Drexel University
Johns Hopkins University
High School, High School at Miller Place High School
Associate of Arts (A.A.), Chinese Studies, Associate of Arts (A.A.), Chinese Studies at Defense Language Institute
Bachelor of Science (BSc), Cybersecurity, Bachelor of Science (BSc), Cybersecurity at University of Maryland Global Campus
English, Chinese