AI Agents

Understand how ai agents mirror travel agents by pursuing goals through actions and tools, using services to check weather, book flights, hotels, and plan activities in an agent loop.

Explore how large language models pair with agents to perform external tasks using tools, with an agent loop that coordinates reasoning and action toward a goal.

Explore how ai agents boost revenue and cut costs by automating tasks, enabling natural language interactions, code assistance, customer service, marketing, sales, and proactive operations.

Explore AI agents concepts and architecture patterns, including autonomy, asynchronicity, and agency, plus memory with LLMs, RAG, tools, and multi-agent collaboration.

Explore Amazon Bedrock, a fully managed, single-api platform hosting multiple LMS in AWS, with security, privacy, guardrails, and tools to compare models, configure prompts, and build knowledge bases.

Set up an AWS account and IAM user with bedrock access. Install AWS CLI, configure your laptop, and install Python with Visual Studio Code to run bedrock labs, noting charges.

Build ai agents: a customer service bot that uses email to fetch purchases and tickets and a weather bot that uses Amazon location service and weather.gov. Explore deployment via sam.

Create a customer service agent in AWS bedrock using the agent builder, define an action group with functions like find customer, find tickets, and totals, and enable code interpreter.

Examine how the customer service agent's lambda function routes events to actions like find_customer_id, find_tickets, or find_purchases via its entry point handler. It unpacks parameters and returns a structured response.

Explore how a weather agent uses external services and two lambda functions in AWS Bedrock to fetch latitude, longitude, and weather forecasts for locations within the United States.

Deploy a running AWS agent with the Serverless Application Model (SAM) and infrastructure as code, by installing AWS and SAM CLIs, deploying templates, testing, and tear-down.

Develop observability and safeguards for AI agents with Amazon Bedrock using CloudWatch; monitor LLM metrics like invocation counts, latency, tokens, and custom guardrail metrics while configuring guardrails and privacy safeguards.

Create and apply guardrails programmatically, check creation status, and handle busy waits. Configure content blocking using topics, content policy, thresholds, and messages for input and output.

Explore why AI projects fail, spot red flags, and use a risk-reward matrix to choose low-risk, high-reward use cases with measurable ROI, data readiness, and stakeholder alignment.