使用 Amazon SageMaker AI 和 Strands 构建和跟踪代理 [重复].pdf

1、 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.A I M 3 7 0-RBuilding and Tracing agents with Amazon SageMaker AI and StrandsKareem Syed-MohammedPrincipal Product Manager TechAmazon Web ServicesDheeraj HegdeSr.Gen

2、AI/ML Geo SSAAmazon Web Services 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.Building Observable,Reliable,and ScalableAgents1.Strands2.Why do we need tracing?3.MLflow for tracing4.Working with traces on MLflow5

3、.Evaluating your agents using MLflow 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.Dheeraj 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.Building Observable and Reliable Agents1.Strands2.Why

4、do we need tracing?3.MLflow for tracing4.Working with traces on MLflow5.Evaluating your agents using MLflow 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.Strands AgentsStrands Agents is an open source python SDK for building agents using just a few lines of code 2025,Amazon Web

5、Services,Inc.or its affiliates.All rights reserved.Why would you choose Strands Agents SDK?FlexibilityEasy tool integration Multi agent collaboration Model flexibility Deployment flexibility 50+and growing list of ready to use tools Easily create my own tools or use Model Context Protocol(MCP)Easy t

6、o put my agents into common multi-agent patterns Native support Agent to Agent(A2A)2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.Strands Agent LoopAgentReasoning(LLM)Tool SelectionTool ExecutionInput and ContextResponse/Outcome 2025,Amazon Web Services,Inc.or its affiliates.All