Skip to content
Search Engineering

Constraints

Each chapter earns one constraint and hands it to the next. Read top to bottom, they trace how a request becomes a governed answer: judgments bound representations, representations bound retrieval, the agent adapts around that evidence, and policy constrains every protected boundary.

Part 1

Search Foundations

  1. Chapter 1 · What Is Relevance?A system optimizes only for the signals that reach it. A metric can rise while the relevance it stands for falls.
  2. Chapter 2 · Corpus and IndexingNo later stage can recover evidence the index discarded; indexing sets the ceiling for every stage that follows.
  3. Chapter 3 · Query UnderstandingA query is an incomplete description of an intent. Reading it well adds evidence about the request without changing the task it expresses.
  4. Chapter 4 · RetrievalBecause lexical and dense retrieval fail on different queries, a trustworthy candidate set needs both.
  5. Chapter 5 · Ranking and FusionRanking is the staged allocation of expensive attention across more candidates than any model can inspect, moving from cheap wide passes to costly narrow ones.
  6. Chapter 6 · GenerationGeneration is bounded by its evidence rather than the model's fluency. Every claim traces to a source, or the answer becomes a refusal.
  7. Chapter 7 · Evaluation and FeedbackBehavioral feedback is biased evidence rather than ground truth, useful only when the full impression is logged and kept apart from interpretation.

Part 2

Agentic Search

  1. Chapter 8 · Agentic OrchestrationAn agent wraps the search system you already run instead of replacing it. Every retry pays for the quality of those tools.
  2. Chapter 9 · Plan and Route Complex SearchesA complex request is a graph of information dependencies, and progress comes from routing each part to a source that can answer it, then replanning from evidence rather than from motion.
  3. Chapter 10 · Maintain State and RecoverAn agent recovers only because its verified progress lives outside the model as structured state; a failure then reruns only the unfinished work.
  4. Chapter 11 · Authorize Agent ActionsInformation, capability, and authority stay separate: retrieving a fact or holding a tool never grants permission to act, and authority comes from deterministic policy rather than model inference.
  5. Chapter 12 · Evaluate Agentic SearchA correct answer reached through a wrong path is still a failure. Evaluation weighs the whole trajectory and its cost, not only the result.

Part 3

Governed Search

  1. Chapter 13 · Permissioned RetrievalAuthorization must shape candidate generation, not merely the final display. Once restricted evidence reaches ranking, caches, counts, or model context, hiding the result cannot undo the disclosure.
  2. Chapter 14 · Policy Through the TrajectoryAuthority is local to each protected decision. Every cache, memory store, delegated service, and resumed checkpoint must verify the current principal, audience, scope, policy version, and validity before using evidence.
  3. Chapter 15 · Connectors and Trust BoundariesA connector can contribute claims, evidence, or effects only under an explicit trust contract. When identity, provenance, freshness, or commit status cannot be established, the trajectory must narrow, reconcile, or stop.