Blog / Security
Posts on Security
Architecting AI Applications for Production
A threat-model-first walkthrough of shipping LLM-backed features — prompt injection defenses, multi-provider failover, structured output, cost control, and the concurrency traps that melt AI services in production.
Zero Trust Security for Microservices
Service-mesh zero trust, in practice: SPIFFE identities, mTLS by default, Istio and Linkerd policy at the sidecar, network policies underneath, and the places where you still need app-level authorization.
Authentication Patterns for Distributed Systems
A threat-model-first walkthrough of JWT, OAuth2/OIDC, refresh rotation, and zero trust auth for Go services — with the mistakes that actually burn people in production.
Authorization Patterns for Go Microservices
Authorization in a service mesh — RBAC, ABAC, policy engines, SPIFFE service identity, and deny-by-default patterns for Go microservices without the panics and silent bypasses.