Skip to Content
View all events

Building AI Applications with AWS Bedrock

Published by O'Reilly Media, Inc.

Intermediate to advanced content levelIntermediate to advanced

Leveraging foundation models for scalable, production-ready AI systems

Course outcomes

  • Deploy and configure foundation models using Amazon Bedrock’s API and Python SDK
  • Build production-ready retrieval-augmented generation (RAG) systems
  • Develop custom AI agents using Amazon Bedrock
  • Implement essential security features in Amazon Bedrock
  • Monitor GenAI applications using CloudWatch

Course description

As organizations move beyond experimenting with AI to implementing enterprise-grade AI solutions, they can turn to Amazon Bedrock, a robust platform for deploying and managing foundation models at scale.

Join expert Eduardo Mota to get step-by-step guidance to leveraging Amazon Bedrock's managed AI services to build enterprise solutions. You’ll learn essential production aspects like implementing RAG systems, managing embeddings at scale, and creating agents that can orchestrate complex business workflows. Using Python and Bedrock’s APIs, you’ll gain hands-on experience with deployment patterns, security configurations, and monitoring solutions specific to AWS’s AI infrastructure. Whether you’re transitioning from other LLM platforms or building new AI solutions on AWS, you’ll gain the practical skills you need to architect and deploy production-grade AI applications using Amazon Bedrock.

What you’ll learn and how you can apply it

  • Deploy and manage foundation models using Amazon Bedrock’s APIs
  • Implement RAG architectures optimized for AWS infrastructure
  • Build custom AI agents that integrate with AWS services and external tools
  • Set up proper monitoring and security controls using AWS best practices
  • Evaluate which Bedrock foundation models best fit specific business needs
  • Design cost-effective architectures for enterprise AI deployment
  • Implement production-ready AI solutions that scale
  • Create secure and compliant AI applications following AWS standards
  • Convert existing LLM applications to run on Amazon Bedrock
  • Build end-to-end AI solutions using Bedrock’s managed services
  • Deploy and manage AI agents for real-world business workflows
  • Troubleshoot common issues in production AI systems

This live event is for you because...

  • You’re a software engineer or developer who wants to move from experimenting with LLMs to building production-grade AI applications.
  • You’re a cloud architect looking to implement secure and scalable GenAI solutions on AWS.
  • You’re a machine learning engineer transitioning to working with managed AI services.
  • You’re a solutions architect responsible for designing enterprise-grade AI applications.
  • You want to build production-ready GenAI applications using Amazon Bedrock.
  • You want to implement secure and scalable RAG systems.
  • You want to create intelligent agents that can handle complex business workflows.

Prerequisites

  • A computer with an AWS account (including access to Bedrock, RDS, Lambda, S3, KMS, and Secret Manager)
  • Access to a Jupyter lab environment such as Sagemaker Notebook or Google Colab
  • Beginner-level familiarity with Python programming and Boto3
  • Basic understanding of REST APIs
  • An understanding of LLM concepts (previous hands-on experience is a plus but not required)

Recommended preparation:

Recommended follow-up:

Schedule

The time frames are only estimates and may vary according to how the class is progressing.

GenAI on AWS (30 minutes)

  • Presentation: Understanding AWS GenAI deployment options
  • Group discussion: When to use a managed service such as Bedrock versus a hands-on service such as Sagemaker Jumpstart?; How do Bedrock and Sagemaker Jumpstart work together in the lifecycle of an AI application?
  • Demonstration: Models available in Bedrock versus Sagemaker Jumpstart
  • Q&A

Getting to know Amazon Bedrock (35 minutes)

  • Presentation: Amazon Bedrock fundamentals
  • Group discussion: How do you know which model to choose?
  • Hands-on exercise: Make simple Bedrock queries
  • Q&A
  • Break

RAG on Bedrock (65 minutes)

  • Presentation: Advanced RAG implementation
  • Group discussion: Does the embedding model really matter?; How can you improve the quality of the models?; How can you integrate with vector databases not supported natively by Bedrock?
  • Hands-on exercise: Set up a RAG system
  • Q&A
  • Break

Agents on Bedrock (65 minutes)

  • Presentation: Building intelligent agents
  • Group discussion: Best practices when designing agentic systems; agents versus RAG—how and when to choose
  • Hands-on exercise: Set up an agent system
  • Q&A
  • Break

LLM security and observability on Bedrock (35 minutes)

  • Presentation: Security with Bedrock Guardrails; observability using CloudWatch
  • Demonstration: Integrating guardrails in the agent system

Wrap-up and Q&A (10 minutes)

Your Instructor

  • Eduardo Mota

    Eduardo Mota is an accomplished AWS Cloud Architect and Machine Learning Specialist. He holds a Bachelor of Business Administration and multiple Machine Learning certifications, demonstrating his relentless pursuit of knowledge. Eduardo's journey includes pivotal roles at DoiT International and AWS, where his expertise in AWS cloud architecture and optimization strategies significantly impacted operational efficiency and cost savings for multiple organizations. Eduardo's commitment to AWS mastery is underscored by his AWS certifications including Machine Learning and Data Analytics, making him a trusted guide in the realm of cloud innovation and artificial intelligence. Beyond his technical prowess, Eduardo's diverse interests, including photography and biographies, reflect his insatiable curiosity and multidimensional perspective.

Skill covered

AWS Lambda