Build and Run Your Own AI Agent in the Cloud

2026-07-01 · Source: Towards Data Science · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure, Software Development & Engineering · Depth: Intermediate, extended

Summary

This article details building and deploying AI agents on AWS using Strands and Amazon Bedrock AgentCore. Strands is an open-source agent framework, similar to LangChain, providing core agent components like LLM integration (e.g., Claude Sonnet 4.6 via Amazon Bedrock), system prompts, and an agent loop. AgentCore offers managed AWS services for agent deployment, scaling, and operational capabilities, including Runtime, Memory, Gateway, and Observability, supporting various frameworks. The demonstration involves creating an educational SME triage agent that answers questions in mathematics, physics, chemistry, and geography. It shows local development, deployment to AWS with AgentCore Runtime, and adding AgentCore Memory to preserve user preferences, such as response style, across different conversation sessions using a `learner-7f83a2` `X-Learner-Id`.

Key takeaway

For AI Engineers or MLOps teams deploying production-grade AI agents on AWS, combining Strands for agent logic with Amazon Bedrock AgentCore provides a scalable and managed solution. You should utilize Strands for defining agent behavior and model integration, while AgentCore handles deployment, scaling, and operational services like long-term memory. Evaluate AgentCore's modular capabilities, such as Memory for user preferences or Observability, to enhance your agent's functionality and ensure robust production readiness.

Key insights

AgentCore provides managed AWS services for deploying and operating framework-agnostic AI agents, complementing agent frameworks like Strands.

Principles

• Agent frameworks define behavior; managed services handle operations.
• LLMs can route complex queries without explicit keyword lists.
• Separate actor and session IDs for long-term memory.

Method

Initialize a Strands agent with a model and system prompt. Use AgentCore CLI to create, develop, and deploy the agent to AWS Runtime. Integrate AgentCore Memory for persistent user preferences via `actor_id` and `session_id`.

In practice

• Use Strands for agent logic, AgentCore for production deployment.
• Implement `X-Learner-Id` for user-specific memory.
• Centralize model configuration for easy LLM switching.

Topics

• AI Agents
• AWS Bedrock AgentCore
• Strands Framework
• LLM Deployment
• Agent Memory
• MLOps

Best for: AI Engineer, Machine Learning Engineer, MLOps Engineer

Related on AIssential

• AWS Bedrock Agents: The Beginner’s Guide — AWS Bedrock Agents provide a managed platform for building and deploying AI agents with integrated orchestration and security.
• Build an AI-Powered Equipment Repair Assistant Using Amazon Bedrock AgentCore — Amazon Bedrock AgentCore enables building RAG-powered AI assistants for complex equipment diagnostics and repair.
• Amazon Bedrock AgentCore harness is now generally available: Go from idea to production-grade agent in minutes — Amazon Bedrock AgentCore harness abstracts complex infrastructure, enabling rapid deployment and management of production-grade LLM agents.
• Semantic search across MCP tools with Amazon Bedrock AgentCore Gateway — Amazon Bedrock AgentCore Gateway enables scalable LLM agents through built-in semantic search for tool selection.
• Amazon Bedrock AgentCore vs Reality — Building reliable AI agents is primarily a software engineering challenge, not just an AI/prompting one.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Towards Data Science.