Title=AWS Engineer Needed – Deploy Containerized Multi-Agent OpenClaw System (Slack + Telegram + E2B + Model Routing) – 1–2 Day Build ⸻ READ THIS CAREFULLY If you have not personally deployed containerized services in AWS (ECS or EC2 + Docker), configured IAM correctly, and implemented Slack OR Telegram bot integrations before — do not apply. Generic AI-written proposals will be ignored. This is a 1–2 day implementation for someone who knows AWS and event-driven systems. ⸻ Project Summary I am building a small internal “AI Office” composed of multiple OpenClaw agents: • Orchestrator • Marketing agent • Research agent • Ops agent Each agent must run as a containerized service in AWS and communicate via SQS (or EventBridge if appropriate). The system must support BOTH: • Slack (Events API) • Telegram Bot API I should be able to communicate with the system from either Slack or Telegram interchangeably. Agents must also be able to spin up and manage E2B sandboxes. This is not a long-term consulting project. This is a focused infrastructure deployment. ⸻ Critical Requirement: Model Routing & Cost Guardrails This is mandatory. I want strict model usage control to prevent runaway OpenAI spend. Requirements: • Default routing to cheaper OpenAI models • Escalation to stronger models only when necessary • Hard token limits per agent • Usage logging per agent • Configurable model selection by role • Simple guardrails to prevent credit burn Preferred provider: OpenAI. If you cannot design model routing logic, do not apply. ⸻ Required Architecture You will: • Deploy secure VPC • Deploy ECS (preferred) or EC2 + Docker Compose • Containerize agents (shared image, role-based config) • Set up SQS queues for inter-agent messaging • Implement Slack Events API integration • Implement Telegram Bot API integration • Secure webhook verification for both • Integrate E2B sandbox API • Store secrets in SSM or Secrets Manager • Configure CloudWatch logging • Ensure internal services are not publicly exposed Slack + Telegram should both route into the same orchestrator. ⸻ Deliverables 1. Fully deployed AWS infrastructure 2. Working Slack bot integration 3. Working Telegram bot integration 4. Agents communicating via SQS 5. E2B integration functional 6. Model routing + cost guardrails implemented 7. Clear README explaining: • Architecture • How to deploy • How to add agents • How to adjust model routing • Where usage limits are enforced ⸻ REQUIRED: Answer These In Your Proposal If you do not answer these specifically, your proposal will be ignored. 1. Would you use ECS or EC2 and why? 2. How would you design model routing to minimize OpenAI spend? 3. How would you implement escalation logic safely? 4. How would you prevent infinite task loops via SQS? 5. How would you validate Slack signing secrets? 6. How would you validate Telegram webhook authenticity? 7. How would you structure IAM roles for each agent? Short, technical answers only. ⸻ Required Experience • AWS ECS or Docker on EC2 • IAM policy design • Slack Events API • Telegram Bot API • OpenAI API usage • SQS / event-driven systems • Secure webhook handling Bonus: • E2B integration experience ⸻ Timeline This is a 1–2 day implementation for someone experienced. If you believe this requires multiple weeks, you are not the right fit.