|
所在平台: Udemy |
课程主页: https://www.udemy.com/course/build-real-world-ai-agents-using-aws-bedrock/
课程评论：没有评论
课程名称：使用AWS Bedrock构建真实世界的端到端AI智能体 课程概述：本课程专为希望使用AWS Bedrock构建生产级真实AI应用的工程师、数据专业人士和软件开发人员设计。您将专注于使用AWS Bedrock、知识库和工作流构建实际工作流，同时利用其他多个AWS云组件，如AWS Lambda、DynamoDB、Redshift和AWS ECS等。您将处理不同领域的真实用例，涉及从检索增强生成（RAG）到完整的多智能体编排等内容。课程采用以代码为中心、可部署的方法，使用核心AWS服务。 您将构建和学习的内容包括： - 使用Bedrock API查询模型，如Claude、Titan和Stable Diffusion - 实现检索增强生成（RAG），使用无服务器的Amazon OpenSearch进行向量搜索，以及Amazon Redshift进行结构性信息的汇聚 - 设计真实的智能应用，通过AWS Lambda调用工具和不同的应用逻辑 - 与DynamoDB和S3进行集成，以自定义逻辑进行数据的提取或写入 - 使用Streamlit构建和部署聊天机器人 - 使用AWS Bedrock设置多智能体协作场景，通过API Gateway使用REST API触发智能体 - 在AWS ECS上通过容器化工作流部署聊天机器人 本课程不涉及理论讲座，而是面向希望将AI系统交付到AWS云基础设施的学员，提供完整的实践示例，覆盖端到端的流程。
This course is designed for engineers, data professionals, and software developers who want to build production-grade and real AI applications using AWS Bedrock. You will focus on building actual workflows using AWS Bedrock, KnowledgeBase and Workflows while leverage several other AWS Cloud components such as AWS Lambda, Dynamodb, Redshift, AWS ECS and many more. You'll work on real-world use cases across different domains covering everything from RAG and tool invocation to full multi-agent orchestration. The course follows a code-first, deployable approach using core AWS services.What you'll build and learn:Use Bedrock APIs to query models like Claude, Titan, and Stable DiffusionImplement Retrieval-Augmented Generation (RAG) using:Amazon OpenSearch serverless for vector searchAmazon Redshift for structured groundingDesign real agentic applications that:Invoke tools and different application logic via AWS LambdaIntegrate with DynamoDB and S3Fetch or write data using custom logicBuild and deploy chatbots using Streamlit Set up multi-agent collaboration scenarios using AWS Bedrock. Trigger agents via REST APIs using API GatewayDeploy chatbots on AWS ECS using containerized workflowsThis course is not about theoretical lectures. It's for people who want to ship AI systems to AWS cloud infrastructure , backed by hands-on examples that work end-to-end.
