Landen Stecker

Security Engineer working at the convergence of operational security, cybersecurity, and AI. CISSP. M.S. Artificial Intelligence at Santa Clara University. Eight years of building and leading security programs in adversarial operational environments. Two years building applied machine learning systems against live adversarial pressure. Three operational chapters across three continents. Bahrain, where I advanced to senior security authority within a single assignment and was selected for the 5th Fleet Admiral's executive protection detail. Poland, where I rebuilt the security program at NATO's newest forward-deployed facility and built a 24/7 GSOC from a compensating control. Norfolk, the world's largest naval station, where I led a 50-person organization with daily executive briefings to the commanding officer. After leaving the Navy in 2023, I turned the same builder instinct toward applied machine learning. The question I set out to test was whether architecture, not resources, could enable a retail practitioner to build infrastructure that performed under continuous adversarial pressure. Crypto was the testbed because it trades 24/7 globally. Active development ran from 2024 through 2025. I built a five-model ensemble coordinated by a central orchestrator: XGBoost and GRU forecasters, a MAML meta-learner, a custom Twitter sentiment transformer, a volume anomaly detector, and a Kelly-criterion risk manager. Uncertainty-aware voting, calibrated confidence scoring, confidence-based abstention, and human-in-the-loop execution. Designed to withstand adversarial data sources from the start. The architectural posture that defended against market manipulation transfers directly to defending AI systems against adversarial inputs, with the threat intelligence stack (CVE, CVSS, MITRE ATT&CK, MITRE ATLAS, OWASP, NIST) as structured training data. That's the research direc