I need an AWS-savvy DevOps engineer who can give our two-person AI/ML startup a solid first pass at production-ready cloud operations without blowing the budget. What I’m after: • A concise blueprint for how to host and auto-scale our OpenAI-powered agents and Streamlit proof-of-concepts on AWS. We currently deploy everything by hand on local machines with Ngrok; I want a basic, reproducible cloud layout that uses EC2, Lambda and S3—plus any other AWS service you feel is the obvious fit. • A lightweight CI/CD pipeline (GitHub → AWS) so each push spins up the latest agent or dashboard, tests it, and, if green, ships it. • Recommendations on repo structure, testing conventions and branching so we improve code quality, shorten deployment time and keep collaboration smooth. Tech context you’ll be plugging into: Python, SQL, OpenAI Agents SDK, Composio, Supabase, Zapier and Intuit Tsheets. We’ll stay on Python unless you can prove Node.ts or React.ts is critical for the front end. Scope is intentionally lean: draft architecture diagram, Terraform or CloudFormation starter files, and a short walk-through call to make sure I can extend the setup myself. If this foundation works, we’ll explore a larger engagement for monitoring, cost optimisation and hardening.