Senior AI Automation Engineer | n8n + Ollama + Mac Studio Agent Builds

Good catch. Stripped the dollar figures and replaced with neutral framing that still conveys "real client work, real stakes" without giving applicants leverage on what you bill.

JOB TITLE: Senior AI Automation Engineer | Build Multi-Agent Swarm on Mac Studio + Subscription Business OS
SALARY: $8 to $10 per hour USD to start, calibrated to demonstrated experience after technical assessment. Senior engineers move toward $15+ per hour as they take full ownership of build delivery.
HOURS: 40 hours per week, full time, long term
SCHEDULE: Overlap with US Eastern Time required (at least 4 hours daily)
TYPE: Direct hire. Not agency, not freelance, not project based.

ABOUT VIRAL GROWTH MEDIA
Viral Growth Media is a Tampa, Florida based digital media and AI automation company. We build full operating systems for high-revenue subscription, creator, and service businesses. We have an in-house Mac Studio M3 Ultra (512GB unified memory, 8TB SSD) ready to host a multi-agent AI swarm on n8n, Ollama, and botctl. We are not hiring you to maintain a running system. We are hiring you to build it.
YOUR TWO PROJECTS
You are not being hired to float across an open queue of small tasks. You are being hired for two named builds with defined scope, defined timelines, and defined deliverables. Both are greenfield. You are building them from zero.
Project 1: Build the Internal Multi-Agent Swarm on Mac Studio M3 Ultra (Greenfield)
The hardware is sitting and ready. The architecture is specified. You are building the system.

Stand up the Mac Studio M3 Ultra as a production inference host. Install and configure Ollama, n8n, and botctl. Lock down system access.
Build the 6-agent swarm from scratch. Each agent has a defined role (content generation, sales triage, monitoring, orchestration, and two more we will brief you on at intake). You will be selecting the right local LLM per agent workload (not just running one default model)
Build inter-agent communication, queue management, retry logic, and persistent memory layers
Build the observability layer: agent uptime, token throughput, queue depth, error rates, escalation triggers
Set up secure remote dev access (Tailscale or Cloudflare Tunnel) so this stack is operable without physical access to the machine
Document the entire swarm architecture so the rest of the team can operate it without you in the room

Project 2: Client Operating System Build (subscription business)
This is an active client engagement currently in proposal stage with a multi-month delivery timeline. You will be the technical lead on delivery. Scope includes:

Helpdesk migration off a legacy phone-based system onto a real ticketing platform (Help Scout, Freshdesk, or Gorgias) with an AI first-touch agent trained on the client's actual three-year support history
Knowledge base extraction and structured documentation
Internal operator dashboard with real-time visibility: active users, trial conversion, failed payments, support volume, app outages, daily revenue
Payment-to-panel middleware between payment processors (Stripe, ---------- , NMI gateways, Easy Pay Direct, Durango) and the client's dashboard layer
Multi-processor failover architecture so a single processor shutdown does not take the business offline
Automated dunning with retry queues and SMS + email recovery sequences
Automated renewal processing, onboarding agent, trial follow-up, cancellation save flow
Customer database migration, deduplication, and clean-up

THE STACK YOU WILL WORK IN
n8n (workflow orchestration), Ollama (local LLM inference on Mac Studio), GoHighLevel (CRM, workflows, IG bot, tags), ActiveCampaign (email deliverability), Whop (community, payments, gated content), Twilio (SMS, A2P compliant), Asana (project management), Google Workspace.
You will be writing JavaScript and Python inside n8n nodes. You will be calling local Ollama endpoints from n8n. You will be building webhook bridges between GHL, ActiveCampaign, Whop, Stripe, and custom panels. You will sometimes be touching raw API calls when n8n nodes are insufficient.
REQUIRED EXPERIENCE

3+ years building automation workflows in n8n, Make (Integromat), or Zapier with measurable production output
Direct experience deploying local LLMs (Ollama, LM Studio, vLLM, or similar). You should be able to tell us which models you have run, on what hardware, and what you learned about prompt latency
Strong JavaScript and Python. You write code, you do not just connect nodes
Direct experience with at least 2 of: GoHighLevel, ActiveCampaign, Stripe API, Twilio API, Whop API
Webhook architecture, REST API integration, JSON debugging, async error handling
English fluent enough to write client-facing documentation. You will be writing SOPs in Asana that paying clients will read

STRONGLY PREFERRED

Experience designing and shipping multi-agent architectures from scratch (CrewAI, AutoGen, LangGraph, or custom n8n agent swarms)
Payment integration experience with US merchant accounts and multi-processor failover architecture
Has built customer-facing AI support agents in production (not demos)
Has worked inside a Mac Studio or Mac Pro environment remotely via SSH or Tailscale
Has set up Tailscale or Cloudflare Tunnel for remote dev access to local infrastructure

WHO YOU ARE NOT
This is not for someone who has only watched n8n tutorials and built one workflow. It is not for someone whose only LLM experience is calling the OpenAI API. It is not for someone who has only ever maintained other people's infrastructure and has never stood one up from scratch. If your portfolio is screenshots of templates instead of production systems you have built and operate, this is not the role.
WHO YOU ARE
You build infrastructure from zero. You read a business and immediately see the leak points. You write clean docs because future-you and the rest of the team will read them. You ship work without being chased.
COMPENSATION AND GROWTH

Starting rate: 13 per hour USD negotiable based on experience, calibrated after technical assessment
Weekly pay
Paid holidays after 90 days
Formal performance reviews at 60 days, 6 months, then annually
Senior engineers move toward $15+ per hour as they take full ownership of build delivery
Direct line to the founder. No middle management layer.

HOW TO APPLY
We screen aggressively. Generic applications go straight to the archive. Follow these steps exactly.

Subject line must read: AGENT SWARM BUILD | [Your Full Name]
In the body of your application, answer the following in this exact order:

One paragraph on the most complex n8n workflow you have built in production. What did it do, what stack did it touch, what broke, and how did you fix it?
One paragraph on local LLM deployment from scratch. Which model, which hardware, which use case, and what was the bottleneck?
List the APIs you have integrated directly (not via no-code tools).
Your hours of availability in US Eastern Time.
Your current monthly billing rate or salary expectation in USD.


Attach or link to:

GitHub or GitLab profile (or equivalent code repo)
At least one screen recording (Loom is fine) of you walking through an n8n workflow you built. Two minutes is enough. Show the actual canvas, not marketing slides.
Any client testimonial or written reference


Do not send a generic resume PDF as your only application.

Applications that do not follow this format will not be reviewed. We get hundreds of applicants per posting. The filter is the test.
We are hiring one person for this role. Project 1 kickoff is immediate on hire. Project 2 kickoff follows shortly after.