Senior AI Agent Infrastructure Engineer (OpenClaw / Hermes)

We are looking for a senior engineer to audit and improve an existing AI agent system built with OpenClaw and Hermes. This is not a prompt engineering role. We need someone who can evaluate the full production stack including agent architecture, LLM usage, cost, memory, security and infrastructure. Initial Scope You will receive SSH access to a Docker environment running OpenClaw. Your first task is to audit the system and deliver a concise report covering: - Agent, skill, hook, cron and dependency structure - LLM model usage and routing - Unnecessary or expensive model calls - Deterministic tasks that should move to Python or Node - Context, memory and session issues - File based state that should move to structured storage - Local versus hosted model usage - Prompt injection and permission risks - Credential and secret handling - Logging, monitoring and reliability - Docker and infrastructure issues For each important finding, include the evidence, severity, impact and recommended fix. Required Experience Strong experience with: - OpenClaw and/or Hermes in production - Python and/or Node.js - LLM model routing and cost optimization - Claude, OpenAI and local open source models - Context and memory management - PostgreSQL or similar structured storage - Skills, hooks, subagents and MCP - Linux, Bash, SSH and Docker - Git, testing and production engineering practices - Prompt injection, least privilege and credential security Experience with Ollama, MLX, llama.cpp, Qwen, Llama and Apple Silicon is highly valuable. Nice to Have - Multi agent fleets with 20+ agents - macOS, Mac Mini or Mac Studio administration - launchd and headless Mac operation - Claude Code or Codex - AWS, GCP or Azure - LLM observability and cost attribution To Apply Please briefly include: 1. A production OpenClaw or Hermes project you worked on 2. One LLM cost optimization example 3. How you decide between deterministic code, local models and hosted models 4. Your experience with agent memory and state storage 5. Your experience with local LLMs 6. Your hourly rate and availability This can grow into ongoing remediation, optimization and fleet operations work after the initial audit.