Jobs → Customer Support → Agentic AI Architect

About the Role We are seeking an Agentic AI Architect to design and build intelligent AI agents capable of reasoning, planning, and managing complex workflows with minimal human intervention. This role requires hands-on experience with AWS AgentCore and Bedrock, as well as expertise in context engineering and harness engineering. Key Responsibilities • Design and build intelligent AI agents capable of reasoning, planning, tool execution, and managing complex workflows with minimal human intervention. • Build and deploy agents using AWS AgentCore, including Runtime, Memory, Gateway, Identity, Policy, and Observability capabilities. • Design effective context management strategies covering memory retrieval, conversation summarization, knowledge injection, prompt versioning, and context window optimization. • Build reliable agent execution frameworks with tool-call orchestration, retry handling, structured output validation, capability routing, and automated testing. • Develop and coordinate specialized AI agents using frameworks such as LangGraph or Strands Agents, including task delegation, state management, and workflow execution. • Build event-driven agents that can monitor relevant activities, identify actionable information, generate notifications, draft communications, and initiate authorized workflows. • Develop knowledge ingestion and retrieval pipelines using PostgreSQL, pgvector, and RAG techniques, with appropriate knowledge governance and access controls. • Build MCP servers and clients to connect agents with enterprise applications, APIs, communication platforms, and internal knowledge systems. • Design and implement solutions using AWS IAM, Secrets Manager, ECR, Aurora PostgreSQL, SQS, EventBridge, and CloudWatch. • Deploy and manage FastAPI and Next.js applications on Amazon EKS using Docker, Kubernetes, Helm, autoscaling, ingress, and AgentCore connectivity. • Develop backend services using Python/FastAPI and frontend applications using Next.js and TypeScript. • Implement enterprise SSO, Cedar/ABAC policies, fine-grained authorization, secure data handling, and protection against unauthorized agent actions. • Implement approval mechanisms and user controls for AI-generated communications, recommendations, and sensitive automated actions. • Build persistent memory, personalized context, conversation continuity, and reliable asynchronous workflows with checkpointing and failure recovery. Requirements • 8-10 years of experience in AI and automation • Hands-on experience with AWS AgentCore and Bedrock • Strong understanding of context engineering and harness engineering • Experience with PostgreSQL, pgvector, and RAG techniques • Knowledge of AWS IAM, Secrets Manager, ECR, Aurora PostgreSQL, SQS, EventBridge, and CloudWatch • Experience with EKS, Docker, Kubernetes, Helm, FastAPI, Next.js, and TypeScript • Familiarity with SSO, Cedar, ABAC, and secure data handling