Patterns / Agentic Systems

Agentic Systems

Foundations 7
Why the Data Layer Is the Prerequisite for AI Agents Prompt Engineering vs Context Engineering The LLM as CPU, Context Window as RAM Why Agents Are Harder Than Chatbots Context Engineering Is the New Product Skill Orchestrator Patterns: Hub-and-Spoke, Hierarchical, Pipeline Coming Stateful vs Stateless Orchestrators — What Actually Matters Coming
Architecture Patterns 6
Fan-Out / Fan-In: Parallelizing Agent Work Coming Designing Idempotent Agent Tasks Coming Context Propagation Without Blowing Token Budgets Coming Circuit Breakers for Unreliable Sub-Agents Agent Pooling: Load Balancing Across Agent Instances Coming Context Isolation: Spawning Agents With Fresh State Coming
Performance & Cost 4
Model Routing: Cheap Models for Routing, Expensive for Reasoning Coming Why Your LangGraph Workflow Is Slow (And How to Fix It) Coming Token Efficiency Across Agent Chains Coming Parallelism vs Sequential — Knowing the Difference Coming
Evaluation & Reliability 5
How the Orchestrator Knows the Sub-Agent Did It Right Coming When More Agents Makes Things Worse Hallucination in Multi-Agent Systems (It's Not Just the Model) Partial Failure Handling and Retry Strategies Observability: Trace IDs Across Agent Chains
Governance & Security 3
What Data Can Your Agent Actually See? Audit Trails for Autonomous Systems Access Control Patterns for Agent Networks Coming
Case Studies 7
AI Agent for Media Planning and Measurement SDR Personalization at Scale Coming Contract Review Across 500 Agreements Coming L1 Ticket Triage Without Hiring Coming Candidate Sourcing Pipeline Coming Monthly Close Automation Coming Inventory and Pricing Decisions at Scale Coming
Stack & Tooling 3
Claude SDK vs LangGraph vs CrewAI — Honest Tradeoffs When Tool Use Is Enough (You Don't Need Agents) Coming Multi-Model Orchestration: Cheap Router, Expensive Executor Coming