Skip to content
ClaudeClaude

AI with Claude on AWS: From code to orchestration

In 45 minutes you'll stand up Claude Code in Amazon Bedrock, teach it your team's conventions with CLAUDE.md, and turn your everyday workflows into reusable agent skills. You'll leave with a working, team-tuned agentic dev environment on Bedrock.

May 20, 202619mWatch on YouTube ↗

At a glance

WHAT IT’S REALLY ABOUT

Using Claude on AWS: secure deployment, tooling, and workshops overview

  1. AWS and Anthropic position their partnership as a long-term, infrastructure-to-model collaboration backed by major mutual investments and purpose-built compute (e.g., Trainium, Project Rainier).
  2. The talk highlights three core benefits of running Claude via AWS—feature-rich platform capabilities, enterprise-grade security/data privacy, and scalable regional deployments for compliance needs.
  3. Amazon Bedrock is presented as a broader developer platform around foundation models, adding evaluation, prompt optimization, fine-tuning, RAG knowledge bases, guardrails, and agentic tooling.
  4. Attendees are shown three ways to use Claude on AWS: directly via Bedrock APIs, via the newly GA “Claude platform on AWS” gateway with feature parity, and via Claude Desktop/Enterprise billing and connectors.
  5. A hands-on workshop walkthrough demonstrates how participants get temporary AWS accounts and a VS Code-like environment to practice Claude Code workflows (Calidraw, MCP/Playwright, Git, sub-agents, hooks, and cost controls).

IDEAS WORTH REMEMBERING

5 ideas

AWS frames “productionizing Claude” as a platform + governance problem, not just prompting.

The session emphasizes surrounding capabilities—observability, IAM/SSO, PrivateLink, SLAs, and compliance alignment—needed to move from local prototypes to reliable cloud operations.

Bedrock’s value proposition is the tooling around models as much as the models themselves.

Beyond Claude access, Bedrock is positioned as a full lifecycle platform: evaluation, prompt optimization, fine-tuning/distillation, knowledge bases for RAG, and guardrails for safety and PII handling.

Security and privacy controls are central selling points for regulated industries.

The speaker stresses keeping traffic within the AWS boundary (no public internet), region choice for sovereignty, and “zero operator access” so neither AWS nor Anthropic staff can access customer inference instances.

You can choose among three Claude-on-AWS integration patterns depending on desired parity and control.

Options include native Bedrock APIs (tight AWS integration), the Claude platform on AWS gateway (feature parity with Anthropic, AWS billing/access control), and Claude Desktop/Enterprise paths with marketplace billing or connectors to Bedrock/platform.

Agentic AI is treated as a first-class cloud workload with hosting patterns on AWS.

Bedrock Agent Core is presented as an infrastructure option compatible with common frameworks (e.g., LangChain) and the Claude Agent SDK, aimed at securely running agents in production.

WORDS WORTH SAVING

5 quotes

As I like to say, we are really just whispering models these days, not really writing code.

Antonio Rodriguez

We avoid this hopping of data going to the public internet. You can keep everything private in your AWS boundary.

Antonio Rodriguez

We have zero operator access, uh, in the platform, which means that no one from Amazon or from Anthropic can actually get access to those instances, so your data remains fully private.

Antonio Rodriguez

I think that if we start diving into the technical part of, uh, this collaboration, there are three angles that we basically offer you when you use Claude through AWS.

Antonio Rodriguez

Amazon has invested, uh, it has done a multi-billion investment in Anthropic, uh, because we really believe that we are building the future of AI together.

Antonio Rodriguez

AWS–Anthropic strategic partnership and investmentsProject Rainier and Trainium chipsAmazon Bedrock capabilities (fine-tuning, eval, RAG, guardrails)Security, privacy, and compliance controls (PrivateLink, zero-operator access)Regional deployment and data sovereignty (e.g., GDPR constraints)Three consumption paths: Bedrock, Claude platform on AWS, Claude Desktop/EnterpriseHands-on workshop: Calidraw repo, MCP/Playwright, Git workflows, sub-agents, hooks, telemetry and rate limits

High quality AI-generated summary created from speaker-labeled transcript.

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.