Engineering at the intersection of AI, security, and research.
I'm an AI and cloud security engineer focused on building systems that stay reliable and defensible outside of clean, single-turn conditions — where behavior compounds, adversaries adapt, and infrastructure has to protect itself by design.
Background
My work sits between applied AI research and production engineering. Through the Air Force Research Laboratory's HBCU/MI Summer Research Program, I've developed behavioral middleware for evaluating concealed intent in multi-turn LLM interactions, and contributed to secure, ephemeral cloud infrastructure for distributed data-querying workloads.
I approach each problem the same way: understand the failure mode first, build the smallest system that closes the gap, and instrument it so the reasoning behind every decision is explainable — not just the output.
I care about engineering that is honest about what it does and doesn't do yet. My project write-ups distinguish implemented capabilities from conceptual future work, because that distinction matters for anyone deciding whether to build on top of it.
Focus areas
Behavioral & Middleware Systems
Multi-turn conversational analysis, intent evaluation, and explainable detection signals.
Cloud-Native Security
AWS infrastructure, IAM, role-based auth, ephemeral compute, and credential-minimized architectures.
Applied Research Engineering
Turning research concepts into benchmarked, documented, testable software.
Principles I build with
Explainability over black boxes
Every detection or decision signal should be traceable to a reason a human can evaluate.
Least privilege by default
Infrastructure should assume compromise and minimize what any single credential or role can reach.
Ship the smallest defensible system
Prove the concept works before generalizing it — then document the boundary clearly.