Job Description & Details
This is a heavy-hitting hybrid engineering role that splits its time between building complex AI agents and setting up the underlying deployment pipelines. If you like the idea of architecting production-grade agentic systems rather than just tweaking prompt templates, this contract gives you a chance to build real things on AWS.
What You'll Actually Be Doing
You will spend about 60% of your time writing Python or TypeScript to build intelligent agent workflows, handle tool calling, and hook up MCP integrations. The other 40% is pure infrastructure workâdeploying those agents onto Amazon Bedrock AgentCore Runtime, containerizing workloads with Docker, and setting up bulletproof CI/CD pipelines in GitHub Actions. Expect to interface with client stakeholders during workshops and meetings, meaning you need to be able to explain technical agent architectures clearly to non-technical folks.
The Core Tech Stack
You can't fake your way through this stack; you need deep, hands-on production experience with the Claude Agent SDK and Amazon Bedrock AgentCore Runtime. This isn't about writing basic prompts in a playground; you need to know how to manage agent runtimes, package applications using Docker and AWS ECR, and lock down cloud security with precise AWS IAM policies. Strong Python or TypeScript fundamentals are non-negotiable since you'll be doing heavy API orchestration and external tool integrations.
Interview Expectations
Expect the hiring team to ask you how you handle infinite loops or unexpected tool-calling failures when an agent goes rogue in a production environment. They want to hear about your battle scars regarding state management and token optimization. Another likely deep-dive will focus on your CI/CD strategy using GitHub Actions for deploying GenAI workloads, specifically how you handle automated evaluation and testing of agent outputs before pushing to AWS Bedrock.
Application Advice
Your resume needs to heavily feature the exact toolsets mentioned in this JD, especially Claude Agent SDK, Amazon Bedrock, and GitHub Actions, to clear the automated ATS filters. Don't just list them as skills; explicitly write bullet points detailing how you architected agentic workflows, managed containerized deployments via Docker, and set up cloud observability. If you have experience with infrastructure-as-code tools like Terraform or AWS CDK, make sure that's visible too.