In this role, you will directly implement AI architectures, deploy production agents, and establish governance frameworks—working closely with cross-functional teams and the client's data leadership to deliver shared AI capabilities on AWS.
Responsibilities
- Design and implement multi-agent systems using Amazon Bedrock AgentCore, including supervisor-collaborator patterns for complex workflows
- Build and manage a centralized Agent Registry for discovering and governing agents, MCP servers, tools, and skills across the organization
- Configure Agent Gateway endpoints to securely connect agents to MCP servers, knowledge bases, internal APIs, and Lambda functions with unified authentication
- Implement agent orchestration using frameworks such as LangChain, OpenAI Agents SDK, Claude Agent SDK, Strands SDK, or custom solutions
- Enable Agent-to-Agent (A2A) interactions and transform legacy REST services into MCP-compatible tools
- Implement OpenTelemetry-compatible tracing via AgentCore Observability to monitor execution paths, debug performance bottlenecks, and audit intermediate outputs
- Configure Amazon CloudWatch metrics for Bedrock runtime monitoring
- Implement model evaluation and testing using Amazon Bedrock Evaluations or Datadog LLM Evals
- Serve as the go-to AI architecture expert for internal teams, translating business requirements into implementable solutions
- Build reusable patterns, templates, and reference architectures for agent development
- Establish standards for prompt engineering, RAG implementations, and agent orchestration
- Mentor team members on AWS AI/ML best practices
Requirements
- Strong hands-on experience with AWS AI/ML services, particularly Amazon Bedrock and its agent ecosystem
- Proven experience designing and implementing multi-agent systems and agentic workflows in production
- Solid understanding of MCP (Model Context Protocol) and agent orchestration frameworks
- Experience with observability tooling such as OpenTelemetry and CloudWatch in AI/ML contexts
- Ability to drive a complex, cross-functional platform initiative as a dedicated, full-time resource
- Strong communication skills for working across technical teams and with stakeholder leadership
Nice to have
- Experience with model evaluation frameworks such as Bedrock Evaluations or Datadog LLM Evals
- Background in platform or enterprise architecture in large-scale organizations
- Familiarity with RAG architectures and semantic data layer design
The assignment is planned to start in Q3, with the possibility of an earlier start during summer. This is a full-time, dedicated consultant role embedded within the client's AI platform initiative.
If you're interested, please apply as soon as possible. We are reviewing applications and arranging client interviews on a rolling basis