Cloud AI Architect

A Cloud Platform Architect with AI workload expertise. The core skills are cloud architecture, infrastructure, and platform engineering. The AI-specific knowledge is focused on understanding workload characteristics — not building models or performing data science.

What this person will do

• Design and deploy cloud-native platforms on AWS — multi-service, auto-scaling, production-grade systems
• Architect microservices that process and structure data at scale with independent scaling, monitoring, and reliability
• Build and manage API gateway / reverse proxy infrastructure for LLM integrations, including routing, audit, cost controls, and access governance across commercial and open-source models
• Expose proprietary databases as secure, governed, AI-accessible interfaces
• Set up container orchestration (EKS/ECS) with auto-scaling policies tuned for bursty, document-heavy workloads
• Design infrastructure for hosting open-source models on-prem or on cloud GPU instances when required
• Prepare infrastructure for model fine-tuning when needed
• Own Infrastructure as Code (IaC), CI/CD, observability, and end-to-end deployment with no handoffs

What this person must bring

Cloud & Infrastructure (Core — Non-Negotiable)

• Hands-on AWS architecture experience (EKS/ECS, VPC, IAM, S3, SQS/SNS, CloudWatch) with real deployments (not just certifications)
• Infrastructure as Code using Terraform or AWS CDK (not click-ops)
• Container orchestration experience with Kubernetes or ECS, including auto-scaling (HPA/KEDA) and service mesh
• API gateway / reverse proxy design with routing, rate limiting, authentication, and audit logging
• End-to-end CI/CD pipeline implementation (fully automated deployments)
• Microservices architecture with event-driven and queue-based decoupling for independent scaling
• Strong networking and security knowledge (VPCs, security groups, encryption in transit/at rest, secrets management)
• Cost management mindset (FinOps, right-sizing, spot vs on-demand decisions)

AI Workload Awareness (Required — Built on Cloud Skills)

• Understanding of LLM API patterns (token-based billing, streaming responses, caching strategies, model routing)
• GPU infrastructure management for hosting or fine-tuning models (instance selection, scheduling, utilization)
• Familiarity with AI tooling (model serving like Triton, vLLM, TGI; vector databases; embedding pipelines)
• Understanding non-deterministic AI outputs and how monitoring/testing differs from traditional APIs
• Experience integrating open-source AI-related libraries into production systems (OCR, NLP, document processing tools)

Operational Mindset (Non-Negotiable)

• Always asks: What happens when this breaks before What's the accuracy
• Treats monitoring, alerting, and reliability as first-class design concerns
• Designs systems for operability, maintainability, and 2 AM debugging
• Comfortable working in regulated environments with audit trails, access controls, and data governance