Cyrus O.

Still doing work by hand that an AI automation system should be handling? Customer messages piling up, social posts taking hours to schedule, five tools that don't talk to each other? I build AI automation systems in n8n, Make and Zapier that fix exactly that, so your business runs smoother without you having to think about it. I'm a full-stack developer who specializes in AI automation and AI agents, not someone who connects three nodes in Zapier and disappears. A few of the systems I've built: • A social publishing system that has posted 456 times to a client's accounts, around 180 a month, with nobody scheduling any of it • A lead pipeline that pulls prospects from four sources into one flow, checks each client's quota before spending on enrichment, and drafts the outreach • A WhatsApp AI agent that books, reschedules and cancels appointments straight into Google Calendar • A market research pipeline that runs five scrapers at once and stops rather than hand over half the data • A chatbot that answers from a company's own documents instead of guessing I design systems that keep working long after the project ends. No fragile setups. No black boxes. No dependency on me to keep it alive. What I automate • AI customer service: emails, DMs, and chat replies handled automatically, with human handoff when it matters • Social media automation: AI captions, hashtags, and descriptions auto-posted across Instagram, TikTok, Facebook, and YouTube Shorts from one upload • Slack and CRM automation: communication flowing into one place with AI summaries, dashboards, and blind-spot detection • n8n workflow automation and AI agents: scheduled, multi-step automations that run your repetitive operations without manual input • Chatbots and voice agents: answering common questions, qualifying leads, handling routine conversations • Zapier and Make to n8n migrations: same automations, better reliability, lower monthly cost, full execution logs • API integration work: connecting CRM, SMS, phone systems, and Google Workspace into one smooth operation AI built into your real tools • Claude, ChatGPT, and OpenAI API integrated directly into your existing workflows • AI agents built with Claude Code and MCP server setup to extend what they can do • RAG systems that answer from your own documents, files and knowledge base • Google Sheets and Workspace agents that read, write, and edit live data for inventory, reporting, and CRM • Voice and phone agents for calls, appointment handling, and routine inquiries When you need more than automation Sometimes the automation needs a real backend behind it. I build that too: Django, FastAPI, and Next.js. The publishing system above runs on a dashboard I built in Next.js and FastAPI, where the client manages every account, schedule and set of credentials themselves. What happens when it breaks Most automations fail quietly. The n8n run finishes green and nothing actually posted. I build the check that catches that, with error branches and alerts, and I document every workflow inline so your team can maintain it without me. If you've got a process that should be running on its own, send me a message with what's eating your team's time and I'll tell you within 24 hours exactly how I'd automate it.