CloudiQS

CloudiQS | London (Remote) We are building production agentic AI systems on AWS that deploy into customer environments within hours. This role is for engineers who understand AWS deeply and want to build AI agents using modern frameworks (LangChain, CrewAI, LangGraph, Strands) while deploying them on production-grade AWS infrastructure. What You'll Build Build agents using modern frameworks (LangChain, LangGraph, CrewAI, Strands Agents) and deploy them on Amazon Bedrock AgentCore Runtime for production scale and reliability. Design agent architectures that leverage AgentCore Memory, Gateway, Identity, Browser, and Code Interpreter services. Implement multi-agent systems using supervisor-worker patterns, agent-to-agent communication (A2A), and Model Context Protocol (MCP) integrations. Build tool integrations through AgentCore Gateway that connect agents to customer AWS services, internal APIs, and third-party systems. Deploy agents as serverless functions or containers depending on execution requirements (Real Time vs long-running). Implement observability using AgentCore traces, CloudWatch metrics, and cost tracking across agent workflows. Own the full stack from agent logic to customer-facing UIs that expose agent capabilities clearly. What We Need Required: Strong AWS Knowledge Production AWS experience (Lambda, ECS, API Gateway, IAM, VPC) Understanding of LLM/agent concepts: tool calling, RAG, reasoning loops, memory management Back End or full-stack engineering (Python primary, TypeScript secondary) Experience with at least one agent framework (LangChain, LangGraph, CrewAI) or willingness to learn quickly Ability to debug distributed systems using AWS tooling Track record shipping and maintaining services in production Strongly Preferred: Amazon Bedrock AgentCore experience (Runtime, Memory, Gateway, Policy) Multi-agent orchestration patterns Event-driven architectures (EventBridge, SQS, SNS) DynamoDB, S3, OpenSearch for agent data storage and retrieval Infrastructure-as-code (CDK, Terraform) Experience with Model Context Protocol (MCP) or API-to-tool conversion What Success Looks Like Agents that fail safely and explain themselves through proper error handling and observability. Tool integrations that are secure, auditable, and properly scoped with IAM. Multi-agent workflows that coordinate reliably across frameworks. Agent deployments that take hours, not weeks. Systems that can be operated by engineers who didn't build them. This is not a research role. This is production engineering building real agents that run in customer AWS accounts. Growth Path Clear ownership of agent architecture and technical leadership across AI-enabled workflows. No artificial ceiling. Interview Process AWS & agent architecture discussion Hands-on coding exercise Agent design session (framework selection, tool design, multi-agent patterns) Production code pairing Ownership conversation CloudiQS is an AWS Advanced Consulting Partner with GenAI, Migration, and Microsoft Workloads competencies.