Adversarial AI testing
Prompt, agent, and tool-chain abuse testing before production rollout.
About me
TX, USA
AI Security Engineer
Jonathan R Reed
I am an AI security engineer focused on testing and hardening LLM systems that need to ship in the real world.
I study Cybersecurity & Risk Management at UT Dallas and work as an EMT-B. That keeps my approach practical, calm under pressure, and grounded in risk prioritization.
Since 2019, I have worked on adversarial AI testing, secure deployment patterns, and product execution for privacy-sensitive environments.
I prefer tight scopes, fast feedback loops, and clear written handoff so teams can ship changes without dragging risk forward.
That bias keeps the work fast, specific, and easier to maintain.
How I work
-
Tight feedback loops
I share progress early so direction stays aligned while we build.
-
Written handoff
Every engagement ends with clear notes your team can execute fast.
-
Private by default
I bias toward local and private-cloud patterns when risk is high.
-
Low meeting overhead
Async-friendly collaboration with concise updates and fast decisions.
My expertise
Core work I deliver for teams building and shipping AI products.
Local LLM deployment
Model setup, quantization, and API integration on your own infrastructure.
Security architecture review
Threat modeling, access control checks, and practical remediation priorities.
Product + security execution
Translate security findings into sprint-ready plans your team can ship.