Learning Path

One curriculum. Multiple section hubs.

Read the site in layers. Start with the foundations, move into control design, then continue into reliability and production thinking. Opinions, tools, and platforms sit beside this path, not inside it.

27 published steps 0 planned next steps Canonical sequence

How To Use This Path

Read in order first

Use the sequence below as the default curriculum. It is the cleanest route through the site.

Use hubs for side depth

Dip into opinions, tools, and platforms when you want lateral exploration without losing the main thread.

Return to the spine

After a side read, come back to the next phase gate so the learning order stays coherent.

Canonical Sequence

Start Here

Frame the discipline before the reader enters the foundations sequence.

Step 1 · Published Introduction to AI Agents: What They Are, How They Work, and When to Use Them Step 2 · Published Why Agent Engineering Is Becoming Its Own Discipline Step 3 · Published What Is Agent Engineering?

Phase 1: Foundations of Agency

Define what an agent is, what changes from plain LLM use, and how autonomy should be understood.

Step 1 · Published What Is an AI Agent? Step 2 · Published LLMs, Workflows, and Agents: What Actually Changes? Step 3 · Published The Autonomy Spectrum: From Stateless Calls to Goal-Directed Systems Step 4 · Published The Sense-Think-Act Loop Step 5 · Published Goals, Constraints, and Success Conditions

Phase 2: Core Mechanics

Show how agent systems decompose work, take action, remember, and reason through multi-step runs.

Step 1 · Published Planning and Task Decomposition Step 2 · Published Tool Use: How Agents Take Action Step 3 · Published Memory: Why Agents Need More Than Context Windows Step 4 · Published Short-Term Context, Retrieval, and Long-Term Memory Step 5 · Published ReAct and the Basic Reasoning Loop

Phase 3: Control and Design Choices

Help readers decide when to add autonomy, how to bound it, and how to orchestrate larger agent systems.

Step 1 · Published When to Use a Workflow Instead of an Agent Step 2 · Published Context Engineering: The New Core Skill Step 3 · Published Structured Outputs, Guardrails, and Execution Boundaries Step 4 · Published Supervisor, Router, and Planner-Executor Patterns Step 5 · Published Human-in-the-Loop Control Design

Phase 4: Reliability and Production Thinking

Move from architecture into evaluation, observability, and the operating discipline behind live systems.

Step 1 · Published Evaluating Agent Trajectories, Not Just Outputs Step 2 · Published Tracing and Observability for Agent Systems Step 3 · Published AgentOps: Running Agents in Production

Phase 5: Reliability Operations and Continuous Quality

Extend production thinking into release gates, ongoing reliability review, and the disciplines that keep live agent systems from quietly degrading.

Step 1 · Published Regression Testing for Agents Step 2 · Published Reliability Reviews for Agents Step 3 · Published Drift, Degradation, and Slow Failure in Long-Lived Agent Systems Step 4 · Published Online Evals vs Offline Evals Step 5 · Published Traces as Test Data: Using Production Runs to Improve Agent Quality Step 6 · Published The Most Common Ways Agents Fail Silently

Parallel Hubs

Opinions

Founder-led arguments, market reads, and point-of-view essays about where agent engineering is actually heading.

Tools

Coverage focused on the surfaces agents touch directly: tool calling, context pipelines, memory, schemas, guardrails, and execution boundaries.

Platforms

Agent SDKs, runtimes, orchestration products, and build surfaces you can use to create custom agents.