IDEAS, DECISIONS & LESSONS

Journal

Notes on AI, architecture and building useful products.

2 articles

16 Mar 2026 · 9 min read

ai-agent

Multi-Agent Deep Dive Part 4: Security Architecture — IAM, Prompt Injection Defense & Audit Logging

Production security patterns for multi-agent AI systems. Covers IAM least privilege per agent, prompt injection defense, input sanitization, output validation, audit logging, secrets management, and compliance requirements.

18 Feb 2026 · 7 min read

aws

AWS Full-Stack Mastery: Security with IAM, Cognito, WAF & Secrets Manager (Part 8 of 10)

Part 8 of 10 — Implement production security for a children's platform: least-privilege IAM policies, Cognito user pools with COPPA-compliant parental consent flows, WAF rule groups, KMS encryption, Secrets Manager auto-rotation, and GuardDuty threat detection.