Engineering essays from the bench.
War stories, architecture decisions, and the occasional tirade. Written by the engineers who shipped it.
Sandbox to Production: What to Do After an Agent Ships
What to do after an AI agent goes live: pin versions, trace runs, cap spend, add a kill switch, grow evals from failures, and alert on behavior.
The Best Tools for Securing AI Agents in Production (2026)
A working stack for securing production AI agents: identity and secrets, gateways and per-tool authorization, guardrail layers, sandboxes, and observability: what each layer does and when you need it.
Vector Databases and Semantic Search: A Practical Guide
What vector search is, how it differs from semantic search, and how to implement it in production: embeddings, indexes (HNSW/IVF), hybrid retrieval, and when a dedicated vector database earns its place.
Securing AI Agents in Production: Threats, Guardrails and Tools (2026)
How to secure AI agents in production: prompt-injection defense, per-tool authorization, least-privilege identities, output filtering, and monitoring, plus the tools that actually enforce it.
AI Agent Hardening: The Production Checklist
A practical AI agent hardening checklist: identity, tool authorization, prompt-injection defense, data boundaries, runtime isolation, and monitoring: what to lock down before an agent touches production.
CVE Slop: The Fake SQLite Bug That Scored 9.8
CVE-2026-51302, a fabricated SQLite use-after-free, scored 9.8 before rejection. How to triage a scary CVE and keep auto-patching safe.
EU AI Act, 2 August 2026: What Changes for Teams Shipping AI Features
Article 50 chatbot disclosure and AI content marking apply 2 August 2026; the Digital Omnibus moves high-risk rules to 2027 and 2028. Engineering checklist.
Vector Search in Postgres: When pgvector Is Enough
Run vector search in Postgres with pgvector: schema, HNSW vs IVFFlat, filtered queries with iterative scans, hybrid search, sizing, and when to move on.
LMDeploy CVE-2026-33626: Your Image URL Field Is an SSRF Endpoint
An LMDeploy image loader SSRF was exploited 12.5 hours after disclosure. How to lock down any LLM endpoint that fetches user URLs, with tested Python.
Model Context Protocol (MCP): Building AI Integrations That Scale
Build MCP servers and clients, connect AI agents to enterprise tools, and deploy Model Context Protocol in production with security and observability.
