Research that can
be checked.

What I’m working toward

I’m building a research practice and looking for paid work where careful investigation matters: AI security, product security, application security, and evidence tooling. The technical examples here are a way to inspect my methods directly.

What I bring to a difficult case

Persistence, HTTP and API analysis, Burp Suite traffic inspection, and a habit of asking what an artifact actually proves. I preserve contradictory records and failed tests instead of editing them out of the story.

My engineering work supports that approach: Security RAG for source-linked retrieval, Cipher for deterministic comparisons, and local role-separated review for challenging interpretations.

How I use AI

AI helps with bounded analysis, code, and independent challenge. It does not approve findings. Evaluators check explicit output contracts; original evidence and human review decide what can be claimed. Method scores and corrective feedback do not train model weights.