Propagate identity and enforce inspection, DLP, routing, endpoint, and response policies at controlled AI boundaries.
Learning outcomes
Curriculum
The chapters are individually addressable documentation pages. You can link directly to a concept from another program, architecture decision, or implementation guide.
Understand the boundary, assets, actors, and policy decisions around AI requests.
Treat the gateway as a controlled decision and evidence boundary, not only a reverse proxy.
Map where trust changes and where data, artifacts, requests, identities, tools, and evidence cross a boundary.
Inventory the data, models, prompts, artifacts, identities, tools, interfaces, infrastructure, and evidence that require protection.
Carry trustworthy identity and permission context through every decision.
Preserve user, workload, service, and tool identity across AI boundaries and make policy decisions explicit.
Preserve actor context while constraining workload, service, and delegated authority.
Enforce source permissions through ingestion, indexing, retrieval, caching, generation, and evidence handling.
Control prompt, file, tool, route, endpoint, egress, and response behavior.
Apply allow, deny, redact, transform, rate, route, and approval decisions to AI traffic.
Make external access, endpoint choice, response inspection, and failure behavior explicit and testable.
Translate sovereignty goals into testable statements of what must be allowed, prevented, approved, observed, retained, and recovered.
Test policies, observe decisions, retain evidence, and control policy change.
Turn acceptance criteria into repeatable promotion decisions with recorded evidence and rollback conditions.
Design signals that explain service health, AI behavior, policy outcomes, change, and incidents without leaking sensitive content.
Design evidence alongside controls so important decisions, releases, configurations, and operating events can be verified.
Version complete AI behavior, assess change impact, approve promotion, support rollback, and retire superseded assets.
Produce and evaluate a gateway policy package.
Test identity-aware allow, deny, redact, route, egress, and approval decisions.
Design exercise; executable implementation pending
Run a synthetic local retrieval system that enforces document permissions, citations, deletion, no-answer behavior, and injection controls.
Runnable synthetic prototype with recorded validation
This curriculum describes policy architecture and testing. It does not provide production DLP rules, guarantee detection of sensitive content, or approve external endpoints. Product integrations and rule packs require separately versioned validation.