【新品发布】利用面向 AWS Analytics 的 AI 代理加速数据工程.pdf

1、 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.Shubham Mehta(he/him)Senior Product ManagerAWS AnalyticsAcceleration data engineering with AI Agen

2、ts for AWS AnalyticsA N T 2 1 5 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.AgendaAgentic AI Strategy for AWS AnalyticsApache Spark upgrade agent for Amazon EMRAmazon SageMaker Data Agent 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.Why working with Data

3、gets harder as you scale?Knowledge GapWorkflow ComplexityCapacity CrunchEvery task spans multiple steps and domainsDeep expertise needed across Spark,SQL,ML.Data grows 10 x every 3 years;teams dontThis is where agentic AI changes the game.2025,Amazon Web Services,Inc.or its affiliates.All rights res

4、erved.2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.Agentic AI in AWS AnalyticsAWS AnalyticsDomain-Specific AgentsMulti-Agent EcosystemAdapts to Your Role/ToolMCP-based Interoperability 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.Apache Spark upgrade agent

5、R EDUC E UPG R ADE T I MEL I NES F R OM MONTHS T O WEEKSAvailable for Amazon EMR on EC2 and EMR Serverless for Python and Scala Spark jobsMCP-based tooling offered through conversational interfaceMulti-step process takes you from planning to code editsError-drive validation loop with data quality gu

6、ardrailsNEW 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.Spark upgrade agent WorkflowMulti-step process with platform-specific adaptationAnalyzes project structure and plansPlanning&Planning&OrchestrationOrchestrationUpdated build config&