Production Notes

I’m a Principal Engineer at StackGen. I write about the systems we build, the mistakes we make, and what we can actually learn from them.

Much of this site is about AI agents: software that uses a language model to choose and carry out steps, often by calling tools. A runtime runs those steps and handles things like tool access and retries. Aiden is StackGen’s platform for running agents for multiple teams. The examples often come from site reliability engineering (SRE): keeping services working and investigating incidents. These are engineering notes from a particular setting, not a claim that every team needs an agent—or needs to build one in Go.

Featured by CNCF: You Can’t Debug What You Can’t See — Observability for AI Agents

Start here

New to the subject? Start with a short reading path rather than the whole archive:

Path Best for
SRE on-call starter pack Triage, RCA, observability — 5 posts
SRE as service starter pack Service Rendered Efficiently — culture + investigation product — 5 posts
Go agent runtime starter pack Runtime definition, Go vs Python, platform split — 5 posts
Full series (searchable) Month-folded archive + search
Service Rendered Efficiently AI investigation as a service product for on-call

Or read these four in order:

  1. What Is an AI Agent Runtime? — plain definition of the production agent loop
  2. Go vs Python for AI Agents — why we chose Go for a production agent runtime
  3. What Are SRE AI Agents? — triage vs RCA vs remediation without demo theater
  4. AI Incident Triage for SREs — What Actually Helps On-Call — what actually helps on-call (on StackGen)

Pocket checklists

Shareable operator checklists (no proprietary schemas):

Topic hubs

Aiden — Enterprise Agent Platform

What Is an AI Agent Runtime? — plain definition of the production agent loop.

AI Agent Runtime vs Platform — Why We Split Them — the engineering trade-off behind turning a single-user CLI agent into a multi-tenant enterprise platform.

Connect