Skip to content

Repository files navigation

Production Notes

Personal engineering blog — enterprise AI agents, workflows, Go, and production systems.

Live at: productionnotes.dev (GitHub Pages source: sks.github.io)

SEO & ops docs: docs/dns-setup.md · docs/search-console.md · docs/distribution-sprint.md · docs/linkedin-carousel-template.md · docs/geo-citation-probes.md · docs/gsc-weekly-checklist.md


Featured series: Building an Enterprise AI Agent Platform in Go

A 30+ post practitioner series on building a production AI agent platform in Go — from language choice through workflows, SRE triage, and enterprise platform lessons. Based on real production work on Aiden, StackGen's enterprise agent orchestration platform.

Series hub: enterprise-ai-agents-go

Starter packs: SRE on-call · Go runtime

Checklists: Evidence-gated RCA · Done?

Distribution: docs/distribution-sprint.md · docs/linkedin-carousel-template.md · docs/geo-citation-probes.md

# Post What it covers
1 Why We Chose Go Go vs Python for AI agents
2 TOML Over YAML Config format trade-offs
3 Go Platform Architecture at Speed Growing a Go codebase fast
4 ReAcTree Bugs Production bugs in agent trees
5 Pensieve Memory Memory for agents that forget
6 Agent Skill Distillation Learning without fine-tuning
7 The HITL Paradox When human approval hurts
8 Defense in Depth Security for tool-wielding agents
9 Observability APM vs agent workloads
10 Terraform Config IaC for agent governance
11 Runtime vs Platform CLI agent → multi-tenant platform
12 Open Source Ecosystem Contributing while shipping commercially
13 JSON Repair Layers Why one repair pass isn't enough
14 LLM Performance Metrics Metrics for agent systems
15 CCE Cloud Entitlements Cloud entitlement patterns
16 AI Incident Triage What helps on-call SREs (canonical on StackGen; /blog/ai-incident-triage-sre/ redirects)
17 Evidence-Gated RCA Multi-plane investigation
18 LLM Tokenomics Context budgets & FinOps
19 Your RCA Agent Needs a Map Runbooks vs topology, gates, verify-first

Later posts in the same series (curiosity gates, completion loops, invent-vs-create, abandon-after-lead, loop salvage, prompt caching, mid-run steering, PII redaction) continue on the series hub.

Topic hubs: workflows · SRE · Go agents


Local development

Jekyll + Minima on GitHub Pages.

bundle install
bundle exec jekyll serve
# http://localhost:4000

Repo structure

├── _config.yml          # Site URL, author, SEO defaults, plugins
├── _posts/              # Blog posts
├── _includes/           # head, header, footer, head-custom, author-bio, subscribe, post-nav
├── _layouts/            # default (skip link + theme JS), post, home
├── _sass/custom.scss    # Dark/light theme tokens + reading UX styles
├── _includes/site.js    # Theme toggle, copy-code, back-to-top (inlined)
├── series/              # Series pillar page
├── topics/              # Topic hub pages (FAQ + schema)
├── tags/                # Tag archive pages
├── assets/images/       # OG social cards (og-*.png)
├── docs/                # DNS, GSC, distribution playbooks
├── CNAME                # productionnotes.dev
└── robots.txt

Writing a new post

_posts/YYYY-MM-DD-post-slug.md
---
layout: post
title: "Your Post Title"
date: YYYY-MM-DD HH:MM:SS -0700
series: "Building an Enterprise AI Agent Platform in Go"  # optional
series_order: N
description: "Keyword-rich one-liner under 155 chars."
tags: [ai-agents, workflows]
image: /assets/images/og-default.png  # or post-specific og-*.png
---

URLs: permalink: /blog/:title/ — no dates in paths.

Publish checklist

  • Confirm the filename uses the publish date and final slug: _posts/YYYY-MM-DD-post-slug.md.
  • Verify front matter before push: title, date, description, and tags are filled in, and add series / series_order when the post belongs to a series.
  • Check whether the default social image is sufficient or whether the post needs its own image value.
  • Proofread the post body for broken links, unfinished placeholders, and formatting issues in headings, lists, and code fences.
  • Make sure the post is ready to publish on its intended date/time, since the existing GitHub Pages workflow will build and deploy from main after push.

License

Content © Sabith K S. All rights reserved.

Connect

About

Personal blog — Lessons learned from building enterprise AI Agentic platform

Resources

Stars

0 stars

Watchers

0 watching

Forks

Contributors

Languages