Role 1 L6
AI/ML/AgentCore
Establish Enterprise Agent Patterns: Define standardized patterns for agent execution, model interaction, tool calling, context management, short- and long-term memory, state/session management, human-in-the-loop controls, retries, resiliency, and blast-radius containment.
Role 2 L6
AI/ML/AgentCore
Enable Framework and Model Flexibility: Establish architecture that can support frameworks such as LangGraph, Strands, OpenAI Agents SDK, Google ADK, and emerging technologies while standardizing enterprise security, governance, observability, and deployment controls. Support MCP and A2A standards where appropriate.
Role 3 L6
AI/ML/AgentCore
Design Secure Tool and Gateway Architecture: Define AgentCore Gateway and MCP patterns for controlled access to enterprise APIs, systems, and tools. Establish identity propagation, authentication, authorization, least privilege, policy enforcement, auditability, external-access boundaries, and Zero Trust principles for agents.
Role 4 L6
AI/ML/AgentCore
Define Memory and Context Architecture: Establish guidance for when information should reside in model context, AgentCore Memory, knowledge bases, vector stores, operational databases, or enterprise systems of record, including isolation, privacy, retention, and lifecycle controls.
Role 5 L6
AI/ML/AgentCore
Establish AgentOps, Evaluation, and Observability: Define standards for tracing, model/tool calls, agent execution paths, token usage, latency, failures, cost, task completion, correctness, faithfulness, safety, tool-selection accuracy, and regression testing. Enable a continuous Observe → Evaluate → Improve → Validate → Deploy lifecycle.
All roles require active AWS certs
