AI Genesis Blueprints

A curated library of production-minded blueprints, templates, and modules for building and operating AI-driven software systems end-to-end.

This repo is not a “cool AI demos” scrapbook. It’s a reference arsenal for shipping real systems: reliable, observable, maintainable, and aligned to business outcomes.

What this repo is

• Blueprints: end-to-end reference architectures (problem → stack → deployment → ops)
• Templates: reusable project starters (API services, workers, UIs, infra)
• Modules: composable building blocks (auth, retries, tracing, RAG, connectors)
• Stacks: known-good combinations you can reach for repeatedly
• Cookbooks: short playbooks and operational guides
• Examples: minimal runnable demos for specific patterns
• Playgrounds: experiments and spikes (kept separate on purpose)

Why it exists

Most “AI app” repos stop at it runs on my laptop.
This one cares about the things that actually matter when users show up:

• failure modes (timeouts, bad inputs, rate limits, partial outages)
• observability (logs, traces, metrics, cost visibility)
• reproducibility (evals, regression tests, prompts/models tracked)
• deployment reality (environments, secrets, migrations)
• business outcomes (time saved, revenue lift, risk reduction)

Repository structure

blueprints/   End-to-end reference architectures
templates/    Project starters (service/worker/UI/infra)
modules/      Composable building blocks
stacks/       Known-good combinations of the above
cookbooks/    Copy/paste playbooks (shipping, ops, evals)
examples/     Minimal runnable examples
playgrounds/  Experiments and spikes
docs/         Philosophy, standards, decision notes, diagrams
scripts/      Repo utilities (scaffolding, checks, etc.)
.github/      GitHub templates and CI