About the course
This two-day, hands-on workshop provides a comprehensive introduction to building and deploying Generative AI applications using AWS Bedrock. We will cover the core service architecture, how to interact with leading Foundation Models (FMs) from providers like Anthropic, Meta, and AI21, and the essential techniques for prompting and retrieval.
You will learn to manage model access, use Bedrock's API via the Boto3 Python SDK, and implement the crucial concept of Retrieval Augmented Generation (RAG) using Knowledge Bases to ground models with your private data. The course culminates in building a fully functional question-answering application.
Instructor-led online and in-house face-to-face options are available - as part of a wider customised training programme, or as a standalone workshop, on-site at your offices or at one of many flexible meeting spaces in the UK and around the World.
-
- Understand Bedrock Architecture: Explain the core components of Bedrock (FMs, Agents, Knowledge Bases) and how to manage access via IAM.
- Access Foundation Models (FMs): Select, configure, and invoke various FMs (e.g., Anthropic Claude, Meta Llama) using the Bedrock API.
- Master Prompt Engineering: Apply best practices for creating effective prompts, handling inputs/outputs, and managing conversation history.
- Implement RAG: Utilize Knowledge Bases and Vector Databases to implement Retrieval Augmented Generation (RAG) for enterprise data.
- Build AI Agents: Understand the role and mechanics of Bedrock Agents for task automation and integrating with company systems.
- Monitor and Deploy: Apply basic monitoring and logging to Bedrock applications and understand deployment considerations.
-
This course is ideal for Developers, Data Scientists, and Solutions Architects who have a working knowledge of the AWS ecosystem and want to start building scalable, production-grade Generative AI applications.
-
Attendees should have:
Intermediate experience with AWS services (IAM, S3, CloudWatch).
Strong proficiency in Python (for using the Boto3 SDK).
A basic understanding of Machine Learning and LLMs is helpful but not mandatory.
We can customise the training for private courses to match your team's experience and needs, for instance including technology primers and more time and coverage of fundamentals for new developers, for instance.
-
This AWS Bedrock course is available for private / custom delivery for your team - as an in-house face-to-face workshop at your location of choice, or as online instructor-led training via MS Teams (or your own preferred platform).
Get in touch to find out how we can deliver tailored training which focuses on your project requirements and learning goals.
-
Introduction to AWS Bedrock
Overview of the Generative AI landscape and the role of Bedrock.
Bedrock's core components: Foundation Models, Agents, and Knowledge Bases.
Security and Access: Managing model access and permissions using IAM roles and policies.
Interacting with Foundation Models
Selecting the right model for the job (e.g., text generation vs. summarization).
Hands-on Lab: Using the Boto3 Python SDK to invoke FMs for basic text generation.
Understanding input/output formats and managing model parameters (temperature, top_p, max_tokens).
Prompt Engineering Fundamentals
Structuring effective prompts: roles, context, and clear instructions.
Techniques for zero-shot, few-shot, and chain-of-thought prompting.
Managing state and conversation history for chat applications.
Data Preparation and Embeddings
Introduction to Vector Databases and their role in Generative AI.
Understanding Embeddings and how Bedrock handles text conversion (e.g., using Amazon Titan Embeddings).
Retrieval Augmented Generation (RAG)
The problem RAG solves (hallucination and proprietary knowledge).
Hands-on Lab: Creating and configuring a Bedrock Knowledge Base (linking S3 data to the vector store).
Implementing the RAG workflow: retrieval, re-ranking, and generation.
Building Task-Specific AI Agents
Introduction to Bedrock Agents and their architecture.
Defining Action Groups (connecting LLMs to Lambda functions for API calls).
Hands-on Lab: Configuring a simple Agent to perform a specific task or query an external API.
Model Customization and Fine-Tuning
When to use prompt engineering versus fine-tuning.
Overview of Bedrock's tools for customization (e.g., using private data for fine-tuning).
Understanding the trade-offs in cost and performance.
Deployment and Monitoring
Integrating Bedrock API calls into full-stack applications (e.g., using Lambda and API Gateway).
Monitoring and logging with CloudWatch for usage and performance tracking.
Best practices for production deployment, scalability, and cost optimization.
-
Core AWS Bedrock Resources
These links are the primary source of truth for the service, its models, and its features:
AWS Bedrock Product Page and Documentation:
The central hub for feature announcements, technical details, and general service overview.
AWS Boto3 SDK Documentation:
This is the official Python SDK that all participants will use to interact with Bedrock's API, invoke models, and manage Knowledge Bases.
https://boto3.amazonaws.com/v1/documentation/api/latest/reference/services/bedrock-runtime.html
AWS Knowledge Bases for Amazon Bedrock:
Documentation detailing how to set up the RAG component, linking S3, and configuring vector stores. Essential for the Day 2 RAG module.
https://docs.aws.amazon.com/bedrock/latest/userguide/knowledge-base.html
Development Tools and Environment
Since this is a Python-heavy course focused on cloud access, these tools are highly recommended:
Python Installation:
Participants need a stable version of Python (3.9+ recommended) installed locally.
AWS CLI (Command Line Interface):
Used for configuring local credentials, interacting with S3, and managing AWS resources outside of the Boto3 script.
VS Code (Visual Studio Code):
The recommended IDE for coding. Its integrated terminal is excellent for managing virtual environments and running Python scripts.
Python Extension for VS Code (Microsoft):
The essential extension for code completion, debugging, and linting when working with Python and the Boto3 SDK.
https://marketplace.visualstudio.com/items?itemName=ms-python.python
Conceptual Resources
For a deeper understanding of the core AI concepts:
Vector Database (Pinecone, ChromaDB, pgvector for PostgreSQL):
While Bedrock manages the vector store, understanding how dedicated vector databases work is crucial for RAG architecture.
Trusted by
