Course Objective & Scope
Taking an enterprise knowledge assistant to production hinges on two requirements: retrieving accurate context from internal documents and reliably blocking responses that leak PII, violate company policy or hallucinate facts outside the source material. This training provides hands-on mastery of Amazon Bedrock Knowledge Bases paired with Amazon Bedrock Guardrails.
Multimodal document parsing, hierarchical and semantic chunking, indexing in OpenSearch Serverless or Aurora pgvector, structured data retrieval, metadata filtering, reranking and Guardrails policies (denied topics, PII redaction, contextual grounding and automated reasoning checks) are practiced hands-on. Participants build a secure RAG reference pipeline and pre-launch safety checklist.
Who Should Attend?
- AI and software engineers building enterprise RAG and assistant applications on AWS
- Cloud architects designing generative AI safety, compliance and governance controls
- Data engineers building knowledge-base ingestion pipelines
- Prerequisite: Python (boto3), basic AWS familiarity (S3, IAM) and core LLM/RAG concepts
Key Takeaways
- Configure Bedrock Knowledge Bases with optimal parsing strategies, chunking modes and vector stores
- Improve retrieval precision using hybrid search, metadata filters, query decomposition and reranker models
- Enforce Bedrock Guardrails for content filters, prompt attacks, denied topics and PII masking
- Block ungrounded hallucinations at runtime using contextual grounding and automated reasoning checks
- Integrate RetrieveAndGenerate citations, streaming guardrail interventions and audit logs into applications
Curriculum
Amazon Bedrock Knowledge Bases Architecture and Ingestion
- Data connectors (S3, Confluence, SharePoint), foundation-model parsing and custom Lambda transformations
- Fixed, hierarchical and semantic chunking strategies; choosing OpenSearch, Aurora pgvector or external stores
Advanced Retrieval: Hybrid Search, Metadata Filtering and Reranking
- Hybrid semantic + lexical search, dynamic metadata filters and implicit filter extraction
- Query decomposition, Cohere/Bedrock rerankers and querying structured SQL data sources
Bedrock Guardrails: Content, Topic and PII Protection
- Content filter thresholds, prompt-attack shields and custom denied topics
- Detecting and masking or blocking built-in PII entities and custom regex patterns
Contextual Grounding and Automated Reasoning Checks
- Catching hallucinations via grounding and relevance thresholds against retrieved chunks
- Verifying responses against formal policy rules with Automated Reasoning checks
Application Integration, Evaluation and Latency/Cost Tuning
- Using ApplyGuardrail standalone or inline with streaming responses and inspecting intervention traces
- Benchmarking RAG and guardrail precision/recall with Bedrock Evaluations
Other Courses in This Category
-
Training, deploying and governing classical ML models and open-weight LLMs on AWS using Amazon SageMaker AI: Unified Studio, JumpStart, Training, Pipelines, Model Registry and Endpoints.
- Live Course
- 5 modules
-
Enterprise AI and agent development on Google Cloud Vertex AI with Gemini models: Google Gen AI SDK, grounding with Vertex AI Search, context caching, safety controls and deployment.
- Live Course
- 5 modules
-
Developing modular multi-agent systems with Google's open-source Agent Development Kit (ADK): LlmAgent, Sequential/Parallel/Loop workflow agents, MCP tools, local evaluation and deployment.
- Live Course
- 5 modules
-
Learn hands-on how to design, deploy, and operate scalable, secure, and managed enterprise agent architectures using Google Cloud Vertex AI Agent Engine and Agent Development Kit (ADK).
- Live Course
- 4 modules
-
Learn how to evaluate, customize, deploy as secure endpoints, and optimize cost-performance for Gemini, partner models, and open-weight models on Vertex AI Model Garden.
- Live Course
- 4 modules
-
Learn how to design and implement end-to-end tracing, telemetry, automated evaluation, safety guardrails, and IAM-based access governance in enterprise cloud agent systems.
- Live Course
- 4 modules
Get in touch
