Field notes
Shipping production agents, in public.
Essays on production LangGraph, agentic RAG, observability, and the realities of architecting agent systems inside large enterprises.
Featured · LangGraph in Production
Gemini Enterprise Agent Platform: What It Is and What You Need to Know
Google's agent stack is four pillars — Build, Scale, Govern, Optimize. Agent Runtime hosts the loop. Studio, ADK, Gateway, and Model Armor are not the same product.
Aug 16, 2026
More writing
Amazon Bedrock AgentCore: What It Is and What You Need to Know
AgentCore is the Bedrock layer for production agents — Runtime, Gateway, Identity, Policy. Agents Classic offered a managed loop; it is closed to new customers.
Aug 15, 2026 · LangGraph in Production
Agent Sprawl: What Happens When Everyone Can Build an Agent
Stripe hit 4,000 no-code agents. Shopify’s Sidekick choked past ~50 tools. Analysts named the pattern agent sprawl. What teams tried, what broke, and what is worth taking away.
Aug 14, 2026 · LangGraph in Production
Docs as Spec: Detecting Documentation Drift with a Claude Agent SDK Agent
How a production docs-quality agent runs guides in a secure sandbox, detects drift, and opens Linear issues plus GitHub PRs — without treating docs as prose.
Aug 3, 2026 · Claude Agent SDK
The 11th-Hour Pivot: Lessons from Scaling AI Observability in the Enterprise
An in-the-trenches story of how an intern-led tracing prototype forced an 11th-hour pivot to LangSmith—and why observability, not model choice, became the true north for production agentic workflows.
Mar 12, 2026 · Agent Evals & Observability
The High Cost of Convenience: Why I Choose Postgres over MongoDB for Production LangChain Checkpointing
A systems-minded take on total cost of ownership and architectural control: why Postgres and PostgresSaver beat MongoDB Atlas for production LangChain agents—and when the 'boring' stack wins.
Feb 28, 2026 · LangGraph in Production
From Black Box to Production Powerhouse: Re-architecting AI Knowledge at Enterprise Scale
How an AI Engineer transformed an opaque internal LLM chat into a production-grade RAG knowledge engine using Postgres, pgvector, LangChain, and Kubernetes.
Feb 25, 2026 · Enterprise & Agentic RAG