ModulesSAI-230
SAI-230 table of contents
Private RAG and Permission-Aware Knowledge
Build grounded retrieval with controlled ingestion, citations, access enforcement, evaluation, lineage, and safe no-answer behavior.
Learning outcomes
What you should be able to do
- Design a private ingestion and retrieval flow
- Implement permission-aware retrieval
- Evaluate grounding, citations, and no-answer behavior
Curriculum
Work through 5 sections in order.
The chapters are individually addressable documentation pages. You can link directly to a concept from another program, architecture decision, or implementation guide.
Knowledge-system foundations
Define the RAG system, its boundaries, and the knowledge lifecycle.
Introduction to private RAG
Design retrieval as a governed knowledge system rather than a prompt attachment.
Define the AI system
Set the intended purpose, actors, assets, dependencies, lifecycle, and system boundary before selecting controls.
Trust boundaries and data flows
Map where trust changes and where data, artifacts, requests, identities, tools, and evidence cross a boundary.
Controlled ingestion and access
Maintain lineage and enforce the source authorization model throughout retrieval.
Controlled ingestion, lineage, and permissions
Preserve source identity, classification, ownership, permission, version, and transformation lineage.
Permission-aware data access
Enforce source permissions through ingestion, indexing, retrieval, caching, generation, and evidence handling.
Identity propagation and policy decisions
Preserve user, workload, service, and tool identity across AI boundaries and make policy decisions explicit.
Retrieval and answer behavior
Design retrieval, reranking, citation, grounding, and no-answer contracts.
Retrieval, reranking, citations, and no-answer
Design retrieval and answer contracts that expose support, uncertainty, and safe refusal behavior.
Grounding evaluation and retrieval threats
Test relevance, support, permissions, injection resistance, stale content, and adversarial cases.
Evaluation and release gates
Turn acceptance criteria into repeatable promotion decisions with recorded evidence and rollback conditions.
Operate and assure
Observe behavior, control change, retain evidence, and record architecture choices.
Observability and operational evidence
Design signals that explain service health, AI behavior, policy outcomes, change, and incidents without leaking sensitive content.
Controlled change and versioning
Version complete AI behavior, assess change impact, approve promotion, support rollback, and retire superseded assets.
Evidence by design
Design evidence alongside controls so important decisions, releases, configurations, and operating events can be verified.
Architecture decision records
Record context, options, control consequences, evidence, dependencies, limitations, and reversal triggers for material decisions.
Apply and assess
Build the evaluation package and verify the module outcomes.
Practical completion package
- Knowledge-source and lineage register
- Permission-aware ingestion and retrieval design
- Grounded answer and citation contract
- Retrieval evaluation dataset and results
- Operational and release evidence package
Current release boundary
Public exercises use synthetic or explicitly approved documents. The curriculum does not authorize indexing customer data or claim that one retrieval stack is suitable for every workload; tested implementation details remain adapter-specific.