AI Automation Engineer (n8n, JavaScript/TypeScript, LLMs)

Hiring from: Europe, LatAm  We're looking for AI automation engineers to help kill repetitive work and ship AI experiments at the company powering outbound for 50,000+ revenue teams. You'll build production automations end-to-end — from n8n workflows to LLM agents to voice bots — working directly with founders and team leads. No PMs, no JIRA, no layers of management. Just fast, high-ownership work wiring together the 20+ tools we run on and turning messy internal processes into clean automated systems. What you'll work on Internal AI automation — n8n flows, LLM agents, voice agents (Vapi/ElevenLabs/Retell), MCP integrations, and glue code across sales, marketing, support, and product. Stack n8n, JavaScript/TypeScript, Python, REST/Webhooks, Anthropic/OpenAI APIs, Vapi/ElevenLabs/Retell, vector DBs, MCP. Must-haves Real coding ability in JavaScript/TypeScript or Python — can debug APIs, handle JSON, and write custom n8n nodes when needed Hands-on n8n experience with non-trivial production workflows Shipped production LLM features — prompts, evals, structured outputs, tool use, retries Reliability mindset — you obsess over error rates, evals, and flows that survive production Integration-heavy background — lots of tools, APIs, bonus for MCP servers Business process thinking — can talk to sales/ops, extract how a process works, and map it into automation Entrepreneurial mindset — ex-founder, ex-agency owner, or ex-consultant preferred Strong written English (async-first team) Based in Europe or Canada with EU timezone overlap Not a fit if you Have no genuine interest in AI or aren't building with LLMs on your own time Need corporate structure — sprints, JIRA, standups, and layers of management Are drag-and-drop only and freeze when you need to write code or debug an API call Are a pure engineer who can't have a coherent conversation about a sales or ops process Get more excited about maintaining existing flows than building new ones Can't clearly explain webhooks, API auth, rate limits, retries, or JSON parsing Can't communicate clearly in English