An engineering ebook · 16 chapters + 2 primers · 7 reading paths
Updated September 6, 2026
AI Systems Orchestration
From Workflows To Autonomous Compound Systems
The frontier is not just bigger models. The frontier is bigger models inside better systems.
Table of contents
Part 0 · The On-Ramp
0aHello, ModelThe gentlest starting point. What a model call really is and why the model remembers nothing.4 min0bFrom Chat To SystemThe bridge from chatting to building. How a model gets hands and what the word agent actually means.4 minPart I · Foundations
00Preface And MapAI systems orchestration is bigger than agents, bigger than workflows, and more durable than any one framework cycle.6 min01Classic OrchestrationThe non AI foundation: DAGs, queues, retries, state machines, durable execution, controllers, and the old systems lessons that still matter.4 min02LLM PrimitivesThe raw building blocks: models, prompts, structured outputs, tool calls, retrieval, context windows, and routing.5 min03Compound AI SystemsThe shift from single model calls to systems: retrieval, code execution, verifiers, sampling, model programs, and feedback loops.4 minPart II · Patterns And Practice
04Agent And Workflow PatternsThe pattern language: workflow vs agent, prompt chaining, routing, parallelization, orchestrator workers, and evaluator optimizer loops.4 min05Context, Tools, And ProtocolsWhere usefulness lives or dies: knowledge sources, context engineering, memory, tool shape, MCP, A2A, Skills, and AGENTS.md.12 min06Runtime OrchestrationWhat happens when the workflow actually runs: graphs, checkpoints, handoffs, parallel agents, human interrupts, and framework choices.4 min07Evals, Guardrails, And SafetyTrust: evals, traces, guardrails, autonomy levels, approval gates, sandboxing, and knowing when not to automate.6 min08Frontier Leverage PatternsThe fun edge: multi sampling, debate, critique, generator verifier loops, search, synthetic data, and spending reasoning where it matters.4 minPart III · Applying And Deciding
09A Real Three-Repo SetupA real three-repo setup where AI agents do actual work: one defines the workflows, one turns requests into plans, and one does the work itself.9 min10Gaps And Project IdeasFrom learning map to possible builds: run ledgers, failure corpora, tool benches, model routers, promotion ladders, and orchestration labs.8 min11Fast Path And Reading ListThe compressed path: build enough taste quickly that you can make good system choices this week.8 min12Stack Selection Field GuideTurning the curriculum into choices: when to use an SDK, graph runtime, durable workflow engine, protocol, or local control plane.11 minPart IV · Operating Agentic Systems
13Learning Loops And Continuous ImprovementHow an AI system gets better after each run without pretending the agent should blindly rewrite itself.8–35 min14Measuring AI-Assisted Software DeliveryA measurement model for deciding whether AI is actually helping software work, not just making more output.10 min15Platform Engineering For Agentic SystemsFrom building agentic systems to operating them as a real platform: catalogs, contracts, evals, governance, and maturity.10 minReading paths
Pick one and the site carries you stop by stop.