On-Site Senior Generative AI Agent Architect

Capgemini; St. Louis (MO); On-site; USD 51,000 - 80,000; Full time

Job summary

Capgemini in St. Louis, MO invites a Senior Generative AI Developer to lead design, development, and deployment of production-grade AI agents using AWS Bedrock AgentCore and Kiro. The role focuses on scalable multi-agent architectures, secure memory management, and robust A2A collaboration.

You will implement agent orchestration, integrate MCP protocols, and build production pipelines in Python/TypeScript within an on-site, contract framework of about 2 months 19 days.

Qualifications

• Experience designing and deploying production-grade AI agents.
• Proficiency with AWS Bedrock AgentCore, Kiro and related services.
• Strong backend skills in Python or TypeScript and AWS services.
• Experience with multi-agent orchestration and A2A patterns.

Responsibilities

• Design Agent Systems: Architect scalable multi-agent systems using Amazon Bedrock AgentCore Runtime and Gateway.
• Implement Collaboration: Build robust Agent-to-Agent (A2A) communication patterns enabling supervisor/sub-agent handoffs.
• Optimize IDE Workflows: Use AWS Kiro IDE and Kiro Powers to turn natural language prompts into production code.
• Integrate Protocols: Connect agents to external enterprise resources, legacy APIs, and tools via the Model Context Protocol (MCP).
• Manage Agent Memory: Build stateful agent sessions using AgentCore conversational memory modules.
• Enforce Cloud Security: Establish deterministic IAM authorization patterns to protect secure microVM agent environments.

Skills

Agile methodologies

A2A patterns

Multi-agent orchestration

Metrics & performance

JIRA usage

ChalkPages

Tools

AWS Bedrock AgentCore

AWS Kiro

Python

TypeScript

JIRA

Job description

Capgemini in St. Louis, MO invites a Senior Generative AI Developer to lead design, development, and deployment of production-grade AI agents using AWS Bedrock AgentCore and Kiro. The role focuses on scalable multi-agent architectures, secure memory management, and robust A2A collaboration.

You will implement agent orchestration, integrate MCP protocols, and build production pipelines in Python/TypeScript within an on-site, contract framework of about 2 months 19 days.