Learn the patterns behind
every AI system
The free, open reference for AI engineering patterns — from RAG and agent orchestration to streaming, fine-tuning, and the classic foundations that power it all. Every pattern ships with diagrams, working Python code, and interview prep.
Whether you're shipping your first AI feature or designing multi-agent systems at scale — this is where you build the mental models that stick.
View all 58 patterns →What you get with every pattern
Every pattern starts with a simple analogy before the technical detail.
Real, runnable snippets using Anthropic, OpenAI, and open-source libraries.
Where teams actually use each pattern — not toy examples.
Curated questions and model answers for every pattern.
Whether you build or learn — this is for you
PatternSimple is designed so a senior AI engineer and a curious beginner both find exactly what they need, without having to wade through what they don't.
Ship production AI systems faster
Stop reinventing the wheel. Learn battle-tested patterns for RAG pipelines, agent orchestration, prompt caching, LLM observability, and streaming — with production-grade Python code you can adapt today.
Browse AI Patterns →Go from zero to AI-fluent
Follow a guided path from the basics to agent systems. Every concept starts with a plain-English analogy and escalates gently to working code. No PhD required — just curiosity.
Start the Learning Path →A learning path that makes sense
Start with the AI patterns powering real products today. Then level up to agent architecture, ML systems, and the software foundations underneath.
Ground AI answers in your own docs using vector retrieval — the backbone of every AI assistant.
Learn →Teach LLMs to reason step-by-step. Simple prompt change, dramatic accuracy boost.
Learn →Give LLMs the ability to call APIs, run code, and interact with the real world.
Learn →Coordinate multiple AI agents — one orchestrator plans, specialized workers execute.
Learn →Stream tokens in real-time for a snappy UX — essential for any chat or generation UI.
Learn →Add safety checks so your agents don't say or do things they shouldn't.
Learn →Centralized ML feature management — share features across models without re-engineering.
Learn →Low-latency online inference at production scale. P99 matters.
Learn →Detect drift, alert on degradation, and keep your model honest over time.
Learn →One AI pattern a day.
Get a short, practical breakdown of one AI engineering pattern every morning — the core idea, a real-world production use case, and a mini challenge. Takes 3 minutes. Builds real fluency.
Today's pattern preview
Agent Loop
Plan → Act → Observe → Repeat. The engine behind every autonomous AI agent and the basis of LangChain Agents.
Explore the full library
58 patterns across AI agents, emerging frontier techniques, and ML systems. Every one is free. Click any to dive in.