Writing

Engineering notes & essays

Thinking through distributed systems, AI infrastructure, and the operational complexity of running AI in production. No tutorials — just engineering analysis.

AI InfrastructureDistributed SystemsPlatform EngineeringObservabilityAI SecurityEngineering Notes

5 posts

All writing

AI Security

The MCP Security Problem: AI's New Supply Chain Crisis

MCP is rapidly becoming the connective layer for modern AI agents. It also introduces an entirely new class of security risks — from prompt injection to tool poisoning to supply chain compromise.

May 202610 min read

AI Infrastructure

Why AI Systems Are Becoming Infrastructure

The shift from AI as a feature to AI as infrastructure is happening faster than most engineering orgs are ready for. Here's what that means architecturally.

May 20268 min read

AI Infrastructure

The Hidden Complexity Behind Production AI

Most teams underestimate what it takes to run AI reliably at scale. Latency budgets, fallback chains, cost controls, and observability are non-negotiable.

Apr 202611 min read

Observability

Observability for AI-Powered Products

Metrics, traces, and logs tell half the story. Instrumenting LLM pipelines requires a new class of signals: semantic drift, token budgets, and prompt telemetry.

Mar 20269 min read

AI Security

Security Considerations for AI Agents

Autonomous agents expand the attack surface in ways that traditional security models weren't built for. Prompt injection, tool misuse, and exfiltration risks need first-class treatment.

Feb 202613 min read