Job Title: AI Automation Engineer (n8n & Voice AI Specialist)

Jobs
24

Hi. I can help with the backend/n8n reliability part of Vapi client builds.

I would start with one paid pilot, not a vague long-term promise: one voice-agent workflow with Vapi custom tools, Supabase/Airtable reads/writes, calendar/CRM edge cases, error handling, and handoff notes so the build can be reused.

My floor for this is $650-$1,000 for the first architecture/reliability slice. Full client builds should be priced by complexity, because a $500 demo is not the same as a production system for a $4k-$10k client.

One question: are your current Vapi tools already designed, or do you need the function schema and n8n data contract created from scratch?

25

Hey @Marco_Lucent,

I build production Voice AI + n8n systems daily — this role is a direct match.

Voice AI — Retell AI (similar architecture to Vapi):

I’ve built and deployed AI Frontdesk agents using Retell AI integrated with n8n workflows — handling real inbound/outbound calls, appointment scheduling via calendar APIs, FAQ handling with RAG pipelines, and human escalation logic. The underlying patterns (custom tools for CRM read-write, function calling, webhook-based state management) transfer directly to Vapi.

n8n — 37 production workflows in active use:

My portfolio covers the full range you’re looking for:

  • Voice AI agent orchestration with RAG + vector DB (Qdrant)
    • Outbound AI calling systems with lead qualification
      • Telegram/WhatsApp AI chatbots with conversation memory
        • GoHighLevel CRM integration pipelines
        • LLM integration: Claude, OpenAI GPT-4o, Google Gemini, Azure OpenAI, and Groq — all wired into n8n with structured output validation, retry logic, and error handling.
      • Tech stack beyond n8n: Python, Node.js, FastAPI, Docker, LangChain, Qdrant vector DB, Supabase.
    • The $500-$1,200 per build + growth path structure works for me. Happy to start with a single pilot build to prove the quality.
  • Best,
  • Ikramul Kayes
  • ikramul.kayes@gmail.com
26

Hi Marco,

I’m interested in this. My background is a close match, though I should be upfront about one thing: I work with ElevenLabs for voice AI, not Vapi. The underlying patterns are the same (custom tools, function calling, calendar/CRM integration via webhooks), but I’d need to get up to speed on Vapi’s specific setup.

Here’s what I bring:

n8n - I run 300+ active workflows in production on a self-hosted Hetzner setup with PostgreSQL, Redis queue mode, and Docker. Error handling, split-in-batches, function nodes, sub-workflows, credential isolation - all battle-tested. I’ve stress-tested builds against real traffic, not just demo scenarios.

Voice AI + Calendar booking - I built a voice agent (ElevenLabs) that handles restaurant bookings end-to-end: checks availability against Google Calendar, asks for guest name, confirms the slot, and writes the booking to PostgreSQL - all through n8n webhooks with custom function calling. The voice agent uses expressive mode with tuned stability/speed settings for natural conversation. If you want, I can walk you through the architecture.

LLMs - Daily work with Claude API (Sonnet for routine tasks, Opus for complex reasoning) and Gemini. Prompt engineering, structured outputs, classification pipelines. I pick models per task based on cost vs quality, not just “use GPT-4 for everything.”

What a workflow screenshot would show - I’d rather walk you through it live than send a static export, since the interesting part is how 300+ workflows interconnect (sub-workflow chains, shared credential pools, error routing, Telegram alerting). Happy to do a 15-minute screen share.

I’m remote from Sweden (CET), available 10-20 hours/week. The project-based compensation model works well for me.

What’s the best way to continue - DM here or do you have a preferred channel?

27

Hi Marco,

This is close to the kind of practical automation work I focus on: n8n workflows that connect real business tools, survive edge cases, and are documented well enough to duplicate.

Where I can help:

  • n8n backend workflows with webhooks, API calls, function/code nodes, retries and error paths
  • LLM/agent steps that return structured outputs instead of loose text
  • OpenClaw-style automation setup and workflow QA
  • handoff docs, test cases, and maintenance notes for retained clients

I have not shipped a Vapi calendar integration as my main proof point, so I would not oversell that. My best starting point would be a paid backend workflow slice or QA/debug task around one client build: map trigger → n8n logic → tool/API calls → logging → test checklist.

If that fits, I can start with a small trial task and show the workflow structure/handoff quality before larger client work.

Best,
Tim

28

Hi,

I’m very interested in the AI Automation Engineer / n8n Voice AI Specialist opportunity.

I specialize in building AI-powered automation systems using:

  • n8n

  • AI agents

  • OpenAI/Gemini integrations

  • APIs & webhooks

  • WhatsApp/SMS workflows

  • Voice automation systems

  • CRM & Google Sheets integrations

  • Multi-step business automations

Over the past few months, I’ve been actively building automation workflows focused on streamlining business operations, client communication, scheduling, and AI-powered workflow orchestration.

Some of the systems I’ve built include:

  • AI assistant workflows

  • Appointment reminder ecosystems

  • WhatsApp & SMS automation systems

  • AI voice and document processing workflows

  • Multi-step follow-up automations

  • Real-time notification systems

I’ve attached workflow snapshots from some of my recent projects, including:

  1. AI-powered Telegram Assistant Workflow

  2. Automated Follow-Up & Reminder Ecosystem

Portfolio:
https://hksautomations.github.io

My focus is on building workflows that are:

  • scalable

  • reliable

  • business-focused

  • easy to maintain

  • optimized for real-world usage

I’m comfortable working remotely, learning new tools quickly, troubleshooting workflows, and building structured automation systems with clean logic and integrations.

I’m currently available and would be happy to discuss how I can contribute to your automation projects.

Looking forward to connecting.

Best regards,
Harvish Shah

Telegram AIBot
Telegram AIBot1466×796 152 KB

Followup scheduler
Followup scheduler906×786 101 KB

29

Hi Marco,

This is a good fit. I build backend automation around n8n/Make, REST APIs, webhooks, CRM/calendar integrations, Google Workspace, Airtable/Supabase and LLM steps.

For your Vapi client builds I would start with one paid build slice: Vapi custom tool → n8n webhook → validation/error handling → calendar/CRM read-write → test script + handoff notes. That gives you a reusable pattern for the $4k-$10k projects before scaling it.

I can work project-based within your $500-$1,200 range depending on scope, and I can start this week once the first milestone/deposit is agreed.

Best,
Carlos Silveira

30

The part I would focus on first is the Vapi tool contract: exactly what the voice agent is allowed to read, what it can write, and which actions need explicit validation before touching a calendar or CRM record. That boundary determines the whole architecture.

For each client build I would deliver: clean Vapi tool definitions with guardrails on bookings and record updates, n8n workflows stress-tested with synthetic call traffic, failed-call and run logging, CRM writeback rules simple enough for a non-technical account manager to follow, and a reusable handoff checklist (env vars, API keys, troubleshooting steps, rollback plan) so you can duplicate the system safely across clients.

Before committing, I want to understand the per-build scope boundary — specifically what counts as the core workflow vs. additional tools or new integrations, and whether the monthly retainer covers bug fixes, new client onboarding, or both.

I have worked with Vapi custom tools, n8n, Supabase, and Google Workspace writeback. Happy to share a workflow architecture note if that helps qualify fit faster, what does your current Vapi-to-calendar flow look like?

31

Hi Marco,

My strongest fit here is the n8n/API backend behind Vapi Custom Tools: webhook design, Code nodes, calendar/CRM read-write actions, Supabase/Airtable state, error branches, retries, structured logs and reusable handoff docs.

For a first paid build, I would suggest one sandbox voice-agent tool path rather than a broad agency commitment: define the Vapi function contract, build the n8n workflow, connect it to a test calendar or CRM, add validation/failure paths, stress-test sample calls, and deliver the documented JSON export plus setup notes.

I do not want to oversell a public Vapi client export I cannot share here, but I can share a sanitized n8n workflow/screenshot privately and build a small proof workflow quickly. First-build pricing would likely fit your $500-$1,200 range depending on how many tools and systems are included; a focused sandbox version should be realistic in 2-4 days once test accounts and scope are ready.