
Explore inference configurations for Bedrock, focusing on temperature, top p, and top k, plus max tokens and stop sequences, to master deterministic versus creative outputs of large language models.
Explore text-to-image generation with diffusion models in the playground by selecting stability AI models in us-west-2, crafting prompts like cat surfing, and understanding model limits and costs.
Set up the code generation project by wiring API gateway to a Lambda that calls Bedrock with a chosen model and saves the output to S3, via Postman.
Code a lambda function and connect it to API gateway for bedrock-based code generation in a serverless AWS workflow. Configure Python 3.11, S3, and CloudWatch logs.
Set up an aws serverless stack for bedrock by creating an s3 bucket, wiring a lambda to a post route in api gateway, and deploying to a dev stage.
Test a live bedrock endpoint with Postman using a Python instruction like implement binary search. Debug via CloudWatch logs and explore lambda boto3 layer updates to enable bedrock runtime.
Create a boto3 lambda layer to address bedrock runtime limitations by packaging the latest boto3 in a Python directory, using a virtual environment, and zipping it for lambda.
Upload a zipped boto3 layer to your lambda function and create the layer with Python 3.11 compatibility. Attach the latest version, deploy, and test for issues.
Learn to test Bedrock models by configuring a Boto3 layer, refining prompt formatting with immediate human text, and deploying with proper status handling, CloudWatch logs, and S3 code output.
Verify Bedrock outputs by examining generated code, including a binary search example with usage. Learn how to return code via API or save it to S3 for AWS workflows.
Set up an AWS Bedrock powered Lambda function to summarize meeting notes from uploaded documents, extracting text from multipart payloads and exposing a Bedrock-based summary route via API Gateway.
Finish building a bedrock-powered lambda function to summarize meeting notes, extracting content from multipart data via Postman and API Gateway, and save the summary to S3.
Create a new post route named meeting summary in API gateway, attach bedrock lambda as its integration in US West 2, deploy to dev, and prepare to test with postman.
Demonstrate invoking the serverless meeting notes summarisation endpoint by posting a txt file to the dev URL using Postman, then debug with CloudWatch until a summary uploads to S3.
Analyze the generated txt file with its UTC timestamp, the original input and its Tesla strategy meeting summary, and note the lambda-driven document chat and stable diffusion project.
Launch a stable diffusion image generation flow with AWS bedrock, API gateway, and a Python lambda. Configure administrator access, apply least privileges, and save images to S3 bucket as PNG.
Update the generative AI workflow to Stable Diffusion 3.5, adjust the payload for text-to-image with a 1:1 aspect ratio and PNG output, and switch to the newer model ID.
Set up a serverless API gateway post route for image generation, attach a lambda to the Bedrock image generation function, deploy to dev, and test with Postman.
Invoke bedrock's stable diffusion model via the image generation endpoint to generate an image from the prompt a cat running a marathon, then save the result to S3.
Set up an automatic evaluation job in bedrock to assess an Anthropic Claude model using accuracy and toxicity metrics, with real toxicity and T-Rex database and S3 outputs.
Evaluate a cheap instant model using toxicity and general knowledge reports, analyzing prompts and responses, and interpreting accuracy scores to assess model performance.
Learn to secure a serverless AWS stack by deleting open endpoints, protect Bedrock integrations with authorization, and distinguish Bedrock from SageMaker while appreciating cost-saving practices.
AWS Bedrock is Amazon Web Services' fully managed platform for building Generative AI applications. It lets you work with powerful foundation models while abstracting away infrastructure complexity. This course will take you from zero to building production-ready AI applications — even if you're completely new to Generative AI.
Projects will integrate AWS Bedrock with API Gateway, AWS Lambda, S3, AgentCore, Postman, and other essential AWS services.
6 Real-World Projects You'll Build
Meeting Summarization — Feed transcripts into a Bedrock-powered pipeline and get structured summaries with key decisions and action items
Code Generation — Describe what you need in plain English, get working code back using foundation models on Bedrock
Image Generation — Generate and control images using diffusion models via AWS Bedrock
Intelligent Multi-Agent System — A hotel & restaurant recommender where Supervisor, Hotel, and Restaurant agents collaborate to deliver personalized results
Voice-Powered Hotel Front Desk AI Agent — Real-time voice AI using AWS Bedrock Nova Sonic that listens, thinks, and responds
Web Search Agent with Memory — A personal assistant built with AWS Strands Agents and Claude that searches the web via DuckDuckGo and remembers users across sessions
What You'll Learn
AWS Bedrock & Generative AI Foundations Understand what AWS Bedrock is, how it fits into the Generative AI ecosystem, and how modern AI applications are built using managed foundation models — explained in a beginner-friendly way.
Large Language Models (LLMs) Learn how LLMs work, what they excel at, and where their limitations lie. Explore real-world use cases including document understanding, chat systems, and code generation using Bedrock-hosted models.
Fine-Tuning & Reinforcement Fine-Tuning (RFT) Go beyond basic prompting. Learn how models can be improved using fine-tuning and RFT with human feedback, when and why to use these techniques, and how they fit into modern AI workflows.
Prompt Management & Optimization Design, manage, version, and optimize prompts for consistent, reliable results. Learn why prompt management becomes critical as applications scale and evolve.
Intelligent Model Routing Discover how to route requests intelligently between different models based on task type, cost, or quality — choosing the right model at the right time.
Batch Inference & Data Automation Process large volumes of data efficiently using batch inference, and automate document and data workflows using AWS Bedrock for scalable AI processing.
Image Generation with Diffusion Models Explore image generation concepts using diffusion models — including how to generate, control, and deploy image-based AI features at scale.
AI Safety, Evaluation & Watermark Detection Evaluate your LLMs for quality, accuracy, and toxicity. Understand AI safety concepts including watermark detection and responsible AI usage.
Serverless Deployment & Scaling Deploy Generative AI applications using a fully serverless stack — AWS Lambda, API Gateway, S3, and Bedrock — and scale AI applications without managing any servers.
AgentCore: From Prototype to Production Take your agents beyond local development. Learn what Amazon Bedrock AgentCore is, deploy agents on AgentCore Runtime, and monitor every tool call and decision with built-in observability — no servers to manage, production-ready from day one.
Agent Memory — Short-Term & Long-Term Give your agents the ability to remember. Implement short-term memory using get_last_k_turns so agents track conversation context within a session. Then add long-term memory with extraction strategies that automatically capture semantic facts, user preferences, and session summaries — so your agents personalize every interaction, even across sessions.
Guardrails & Responsible AI Learn how to use Amazon Bedrock Guardrails to add safety controls, filter harmful content, reduce risk, and enforce responsible AI behavior across your applications.
Model Comparison Compare different foundation models based on quality, speed, cost, capabilities, and use case fit so you can choose the right model for each task.
AWS Model Usage Quotas Understand AWS Bedrock model usage quotas, request limits, throttling, and how quotas affect real-world production applications.
Inference Profiles Learn how inference profiles work in Amazon Bedrock and how they help with scalable, reliable, and cross-region model invocation.
Marketplace Models on AWS Bedrock Explore Amazon Bedrock Marketplace Models and learn how to access additional foundation models beyond the default Bedrock catalog.
Advanced Bedrock Features Go beyond basic prompting and deployment by working with more production-focused AWS Bedrock capabilities used in real AI systems.
Understanding the AWS agent harness and runtime architecture for various multi-agentic workflows
Swarm-based agent collaboration
Graph-based workflows
Agents as Tools
Agent-to-Agent communication
By the End of This Course, You Will Be Able To:
Build and deploy Generative AI applications using AWS Bedrock
Summarize documents and extract key decisions with AI pipelines
Generate text, code, and images using foundation models
Improve model behaviour with fine-tuning and reinforcement fine-tuning
Manage and optimize prompts at scale
Route requests intelligently between models
Choosing the right multi-agent pattern for your use case
Run batch inference jobs for large datasets
Evaluate and monitor LLM outputs for safety and quality
Deploy scalable, serverless AI systems on AWS
Build intelligent multi-agent AI systems
Create real-time voice AI agents using AWS Bedrock Nova Sonic
Deploy agents to production on AgentCore Runtime
Add short-term and long-term memory to agents
Build web search agents with Strands, Claude Haiku, and DuckDuckGo
Who Is This Course For?
Aspiring AI developers who want practical, hands-on skills with AWS
Cloud engineers looking to add Generative AI to their toolkit
Curious beginners who want to understand and build with modern AI — no prior experience required
Whether you're new to AI or an experienced cloud engineer, this course gives you practical skills, real projects, and a strong foundation for working with AWS Bedrock and modern AI systems.
Join us and explore the future of Generative AI with AWS Bedrock — from prompts to production.