Taylor Morgan

Building SecureStack AI taught me that grounding an LLM's answers is the easy part. The real work was a verification layer that checks citations, tool-call evidence, and deterministic scanners so nothing gets trusted just because it sounds confident, plus two AWS deployment gotchas along the way.

I Built an AI Tool to Review Code for Security Issues. Here's What It Took to Make It Trustworthy

The account-level decisions behind deploying SecureStack AI to real ECS and RDS infrastructure: a scoped IAM identity instead of root, two separate permissions boundaries to close a subtle PassRole gap, a compensating IP restriction for a longer-lived demo, and what it actually cost.

Under construction 🚧