Hiring : AI Automation Engineer / n8n & AI Agent Developer

Jobs
23

Hi,

This caught my attention because it describes my actual day-to-day work, not a wish list.

I build and maintain a production SaaS platform for the restaurant industry, entirely on self-hosted n8n + PostgreSQL. Right now it runs 300+ active workflows on a Hetzner VPS with Redis queue mode, Docker Compose, and 163 row-level security policies for multi-tenant data isolation. It’s been running unattended since late 2024.

Here’s what maps directly to your requirements:

n8n workflow architecture - I build with proper error handling, retry logic, dead-letter patterns, sub-workflows, and credential isolation. Not the “connect three nodes and hope” kind. Every workflow is documented and another developer could pick it up.

AI/LLM integration - 30+ agents orchestrated via Claude API and Gemini. Structured outputs, function calling, classification pipelines, prompt engineering. I use different models for different tasks (heavier models for reasoning, lighter for routine classification).

Document processing - I’ve built OCR and PDF pipelines for invoice processing, contract extraction, and receipt scanning with AI classification.

CRM and business system integration - Google Workspace (Drive, Gmail, Calendar, Sheets), REST APIs, webhooks, PostgreSQL as system of record.

Browser automation - Playwright and Chrome DevTools for scraping and interacting with systems that don’t have proper APIs. Built lead-gen pipelines that process 2000+ businesses per run.

Human-in-the-loop - Button-click approvals, Telegram notifications, email digests. I never auto-send anything that touches a customer without a human check.

Background: 24 years in hospitality operations before moving into tech. That means I understand messy business processes from the inside, not just from a spec document. I know what it’s like when the system needs to work at 11pm on a Friday.

I’m based in Sweden (CET), available 10-20 hours/week remote, and prefer project-based or retainer arrangements. Happy to discuss scope and pricing for a test project.

DM me or reply here - I’d like to hear more about the types of clients you’re working with.

24

Hey Tomas,

Your line — “AI does not replace architecture” — is exactly right, and it’s the gap I see in most AI automation projects.

I run noorflows, a solo n8n consultancy focused on production-readiness. My background is 18 months of algorithmic trading infrastructure in Python (multi-process orchestration, retry-with-backoff, audit logging) — the same patterns that make document processing, CRM integrations, and AI agent workflows survive past week one in production.

On your stack specifically:

  • n8n + AI agents: I build with idempotent webhook handlers, confidence-score routing (classify → route → escalate on low confidence), and human-in-the-loop checkpoints as first-class workflow nodes, not afterthoughts.
  • Document processing / OCR: n8n → external OCR (Vision/Textract) → structured extraction → validation loop → database write — with dead-letter queues for unreadable documents instead of silent failures.
  • Browser automation for legacy systems: I pair Puppeteer/Playwright with n8n’s Execute Command or HTTP nodes, adding retry logic and screenshot-on-failure for debugging legacy integrations that break without warning.

Every build follows a 6-dimension production-readiness framework I published here on the forum: The 6-dimension production-readiness checklist ( The 6-dimension production-readiness checklist I've been using on every n8n workflow review )

I work async-first, remote, project-scoped with fixed pricing. Happy to start with a scoped pilot on one of your active client projects so you can evaluate before committing to a longer partnership.

Portfolio + details: noorflows.com

25

Hi Tomas,

Your line “AI does not replace architecture” — that’s exactly the problem I solve for clients.

What I build (production, not demos):

-> n8n + Claude AI pipelines: classification, extraction, human-in-the-loop validation before action

-> Error handling as architecture: Try/Catch branching, exponential backoff on rate limits, dead-letter queues, Slack/email alerts on failure — no silent failures

-> Document processing: OCR (Google Vision/Textract) + structured LLM extraction + validation loop

-> API integrations: Google Workspace, REST APIs, Supabase/PostgreSQL, CRM systems

-> Browser automation: Puppeteer/Playwright for legacy systems without APIs

In production (4+ months, zero interventions): Inbound lead pipeline — Webhook → Claude qualification → CRM write → team notification. Fully unattended, every execution logged and auditable.

Stack match: n8n · Claude/Anthropic API · JavaScript · Python · Supabase · PostgreSQL · Docker · Make.com

Available immediately for project-based remote cooperation. EU timezone, async-first.

Sending formal application with workflow documentation to info@s-e.lt.

Oktay | Flowmatic Works

flowmatic.works@gmail.com

26

Hi @Tomas_Maciulskas,

Your post resonates because you’re describing exactly how I work: take a messy business process, design the architecture, build it in n8n, and deliver something that runs reliably in production without hand-holding.

What I bring to the table (production systems, not demos):

  • 37 production n8n workflows currently running across two self-hosted instances: Voice AI agents (Retell AI), RAG chatbots with Qdrant vector DB, CRM integrations (GoHighLevel), cold outreach pipelines (Apify + AI personalization), Telegram/WhatsApp AI bots, and content automation
    • LLM orchestration in n8n: Claude, OpenAI GPT-4o, Google Gemini, Azure OpenAI, Groq – structured output validation, function calling, classification pipelines, retry logic on every workflow
      • Python + JavaScript: FastAPI backends, LangChain RAG pipelines, custom n8n code nodes for data transformation and validation logic
        • Browser automation: Apify actors + Playwright for scraping and interacting with systems without APIs
          • CRM/ERP integrations: GoHighLevel, REST APIs, webhooks, Supabase, PostgreSQL
            • Error handling as architecture: retry logic, dead-letter queues, Slack/Telegram alerts on failure, structured logging – nothing fails silently
          • Portfolio with screenshots of all 37 workflows: GitHub - ikramulkayes/n8n-portfolio · GitHub
        • I’m based in Canada (EST), available for project-based remote cooperation, async-first. Happy to start with a single scoped pilot to demonstrate delivery quality.
      • Also sending a formal application to info@s-e.lt.
    • Ikramul Kayes
    • ikramul.kayes@gmail.com
28

Hi,

Your requirement strongly matches the type of automation systems I am currently building using n8n and AI integrations.

I specialize in creating AI-powered automation workflows that connect business operations, communication systems, APIs, and AI agents into one streamlined ecosystem.

My experience includes:

  • n8n workflow development

  • AI agent orchestration

  • OpenAI/Gemini integrations

  • WhatsApp & SMS automation

  • CRM and Google Sheets automation

  • Webhooks & API integrations

  • Scheduling and follow-up systems

  • Voice and document processing workflows

Recently, I worked on automations involving:

  • AI assistants

  • Automated client communication systems

  • Lead and appointment workflows

  • Real-time notification pipelines

  • Multi-step automation logic

I enjoy building workflows that are not only functional, but also clean, scalable, and easy to expand in the future.

I am comfortable working independently, understanding business requirements quickly, and delivering production-ready automations with proper structure and reliability.

Would be happy to discuss your project further.

Best regards,
Harvish Shah

29

Hi Tomas.

What attracted me to this post is that it focuses on solving business problems, not just connecting automation nodes.

My background combines consulting, process analysis and AI automation. I am used to understanding how a business process works, identifying bottlenecks, defining the target workflow and then implementing the automation.

Relevant experience includes:

  • n8n workflow design and implementation.

  • OpenAI and LLM-based automation.

  • REST APIs, webhooks and third-party integrations.

  • CRM, email, calendar and document workflows.

  • Human-in-the-loop approval processes.

  • RAG assistants and business knowledge bases.

  • Python and JavaScript for custom logic when required.

I am comfortable taking ownership from requirements gathering through delivery, testing and documentation.

I have also been experimenting with OpenClaw-style agents running locally and integrating them with external tools and automation workflows.

One area where I believe I add value is translating business requirements into practical automations. Many projects fail because the process itself is not clearly defined before the technology is implemented.

Availability: remote and project-based.

Happy to discuss current projects and see whether there is a good fit.

Juan Antonio Molina

30

Hi, your post lines up well with what I build. For about a year I’ve been working in n8n and AI automations, including AI agent style flows. Two things I’ve shipped: a WhatsApp assistant that lets me run Gmail straight from chat (read, draft, send), and a receipt-to-Sheets pipeline that uses Gemini to pull details off a photo and log them on its own. Inside n8n I’m comfortable with APIs, webhooks, and LLM calls, plus custom logic in JS Code nodes and Python where it helps. You can see everything here: SakoA10 (Sako Andonian) · GitHub . Two things I like to check early: is the role open to someone based in Lebanon, and how do you handle pay for remote contractors?

31

Hi there,

I’m Niklas, an AI Automation Engineer based in Berlin with 10+ years of software dev experience. I specialize in exactly what you’re looking for — n8n, AI agents, RPA, and business process automation.

Highlights:

  • Multi-agent document processing with n8n + LangChain + LLMs

  • ERP, CRM, logistics system integration via REST APIs

  • Docker, Python, TypeScript for custom nodes

  • Enterprise & SME automation delivered

Your “real business processes” focus aligns perfectly with my work — I’ve built enterprise AI agent platforms (Dify + n8n) for insurance and e-commerce chatbots.

Profile: https://freelancermap.de/freelancer/niklas-schmidt

Looking forward to discussing how I can help Synergy Effect!
— Niklas

32

Hi Tomas / Synergy Effect team,

I’m Kevin. I saw this role and the direction caught my attention because it is focused on real business workflows rather than simple demos.

I want to be transparent about fit: I would not present myself as a senior automation architect with years of production n8n ownership. My stronger fit is as a practical AI-assisted automation builder who can help with scoped workflow slices, research-to-prototype work, Google Workspace / email / Sheets workflows, API integration, Python/JavaScript support, and clear documentation.

For a first collaboration, I would prefer a small paid mini-task or solution design exercise, for example:

  • classify and route incoming emails with human approval;
  • extract structured fields from documents or messages;
  • write clean workflow logic with fallback/error cases;
  • prepare a test plan and handoff note for a workflow slice.

My current availability is about 10 hours per week. For project-based work I would expect around USD 30/hour, or a fixed price for a well-scoped mini-task.

I can share a short visual workflow sample showing how I structure an AI-assisted Google Workspace workflow with review points and handoff notes. If this level of support could be useful alongside your senior builders, I would be happy to do a practical test.

Best,
Kevin

33

Hey Tomas — this looks right up my alley. I just emailed my details to info@s-e.lt too, but figured I’d drop a note here in case this is where you’re looking.

I build AI agents and automations on n8n — it’s pretty much all I do. Some of what I’ve built: a customer-facing agent that handles messages across email, text, Telegram and Facebook and actually remembers the conversation; a setup that pulls in documents (PDFs, video, whatever) and makes them searchable for the AI; and a lead-finder that scrapes businesses, scores them with AI, and sorts them automatically.

Honestly most of my time goes into the boring-but-important stuff — making sure things don’t quietly break, catching the weird edge cases, handling errors. Your line about AI not replacing real architecture hit home.

I haven’t shipped a fully autonomous agent yet (played with it locally, but my real work’s all n8n so far) — but I’d love to get deeper into that.

Available now for project work. Here’s my portfolio if you want to see the actual builds: GitHub - jakegilbert-dev/ai-automation-portfolio: Production AI agent systems built on n8n — case studies and sanitized workflows · GitHub

Cheers,
Jake

34

Tomas, I’m Misha, an n8n and AI automation developer. I read your post and it matches exactly how I work — I don’t just wire nodes, I design production-ready agentic flows that handle document processing, CRM/ERP integration, and browser automation with proper error handling and retries. I’d build reliable workflows using n8n combined with Python for OCR structuring and RAG logic, integrating systems via REST APIs and webhooks while keeping validation, logging, and human-in-the-loop approval steps baked in from the start. I’ve delivered similar projects involving invoice automation, email classification, and multi-system data routing where the agent doesn’t just guess — it follows structured process logic. If you describe a current process that needs automation, I can estimate and deliver it fast. Portfolio: https://mikedevai.netlify.app/ | Telegram: Telegram: Contact @hely_chatbots | WhatsApp: +375293761570

35

Hi Tomas,

Practical n8n is my core, so this fits.

What I built: the full lead gen and follow up backend for a marketing agency. n8n orchestrating GoHighLevel, Twilio, and the Anthropic API, with AI agents that qualify and route leads and generate personalized outreach.

Systems integrated: GoHighLevel (CRM), Twilio, Instantly, Apify scraping, Anthropic and OpenAI APIs, webhooks and REST across the stack.

Business problem: client onboarding and follow up were manual and leads slipped through the cracks, which capped the agency’s growth.

Solution: production n8n workflows with branching, retries, error handling, alerting, and human review where needed, plus AI agents that classify, extract, and route data, so nothing runs blind.

Result: 2,000+ messages a month with zero manual input, send volume scaled 5x, and failures surface instead of silently dropping leads.

OpenClaw: I have not used OpenClaw specifically, but I build agentic n8n workflows where agents use tools, call APIs, and act on data with validation and approval steps, not just generate text.

Availability: open now for project based remote work, comfortable async, US Eastern hours.

Cooperation model: project based to start, open to a longer partnership. Around 50 USD an hour, or fixed per scoped project.

Happy to jump on a short intro call. What is the first business process you want to automate?

Luke

36

Is there a compensation structure or some sort of pay rate?

37

Useful way to scope this before hiring long-term: pick one workflow that has a clear input, one external system boundary, one approval checkpoint, and a visible success receipt.

For an AI automation builder, I’d ask candidates to show the run log, failure handling, and the exact human handoff point, not just a demo video.

I’m validating a 72-hour agent job sprint for exactly this kind of work: one narrow workflow, clear proof receipt, and a hard out-of-bounds list before anyone commits to a bigger engagement. If that would be useful, the first sprint I’d suggest is a messy-process teardown: intake doc/email → classify/extract → human approval → CRM or sheet update, with a receipt showing every decision and failure path.

The question I’d ask before scoping it: which client workflow creates the most manual rework today, and which irreversible actions must stay human-only?

38

Varsha T S — AI Automation Engineer this is my work

39

One thing I’d add to this hiring spec: the strongest candidates should be able to show a failure-mode map for one real workflow, not just a demo.

For an AI agent / n8n role like this, I’d ask them to walk through:

1. what happens when the model returns uncertain or malformed output

2. where human approval is required before an action touches CRM, finance, or a customer

3. how retries, fallbacks, and alerts are handled

4. what data is logged for debugging without leaking sensitive client data

5. how the workflow can be handed over so another operator can safely maintain it later

That usually separates production automation people from “I connected a few AI nodes” applicants. A small paid design exercise around one messy document/email/CRM workflow would probably reveal more than a generic portfolio review.

40

For this kind of production automation role, I’d make the hiring test a small paid workflow with a required closeout receipt. The receipt should list: systems touched, API/browser actions taken, test cases run, known exclusions, failure handling, and rollback steps. That gives you a much better signal than a portfolio demo because it reveals whether the developer thinks about operational ownership. In particular, browser flows and API-triggered mutations need visible verification before the work is called finished. EloPhanto’s completed-work signature receipt is relevant as a sample format for making “done” auditable instead of subjective.

41

Hi Synergy Effect team,

Your AI Automation Engineer posting maps almost exactly onto what I already do day to day: build AI workflows and system integrations for small businesses.

A few things I’ve shipped solo, end to end:

  • autoFacture (autofacture.pro) — a compliant French e-invoicing SaaS: integrates the Hiboutik POS API + LemonSqueezy payments and generates standards-compliant (Factur-X) invoices. Next.js / TypeScript / Supabase.
  • A POS → producer stock-management tool bridging Hiboutik with automated vendor communication (OpenAPI-driven monorepo).
  • An AI feature (Anthropic Claude) wired into a live client site.

I work AI-native — I ship fast with modern agentic tooling — and I’m comfortable talking to non-technical clients, which I do constantly.

I’d love to hear how your engagements are shaped and whether there’s a fit. Happy to walk through any of the above live or take on a practical task. Live to click: autofacture.pro · my studio + client sites: sitetable.dev.

Best,
Robert Turner · info@robertturner.co · sitetable.dev

42

This lines up well with what I actually build. I run production AI-automation for exactly these messy processes: document/OCR extraction, email and inbox triage, CRM/ERP integrations, and agents that classify, extract, validate and route business data instead of just wiring a few nodes.

Two things I’ve shipped that map directly:

- Licitagram: an autonomous system where 40 AI and deterministic workers ingest public tenders, run OCR and semantic extraction, score and classify each one, and an LLM lets users query the document directly. It runs 24/7 in production on n8n, Python, Postgres and LLM APIs.

- Fidacy: a firewall for AI-agent actions. The agent classifies and validates each action against a policy, denies anything out of scope, and every decision is logged and independently verifiable. That is the classify/extract/validate/route/act loop you describe, built for a high-stakes case (payments).

I’m comfortable taking a vague business process, asking the right questions, scoping and estimating it, then building, testing and delivering something reliable rather than a demo. Stack: n8n, Python/Node, browser automation, OCR, and the usual LLM/agent tooling.

Closest single link: https://fidacy.com

Available now, remote, based in Lisbon (UTC+1). What’s the first process you’d want scoped as a paid pilot, and do you work fixed-price per automation or hourly?

43

As a Software Architect and Backend Engineer, I specialize in designing scalable software solutions, enterprise automation platforms, and AI-powered systems. My experience spans software architecture, backend development, distributed systems, and intelligent automation, helping organizations modernize business processes and integrate complex enterprise environments.

My primary expertise is in Python, complemented by JavaScript, Perl. I design and develop robust backend services, REST APIs, microservices, and enterprise integrations using frameworks such as FastAPI, Django, and Flask.

I have extensive experience building production-grade automation platforms using n8n, Robot Framework, ServiceNow, and Selenium, delivering scalable workflows that automate business processes, integrate heterogeneous systems, and improve operational efficiency.

My work also focuses on Artificial Intelligence, designing AI agents and LLM-powered solutions using OpenAI, Google Gemini, Vertex AI, and Retrieval-Augmented Generation (RAG). I develop intelligent workflows for document processing, conversational AI, enterprise knowledge retrieval, and business automation.

I design event-driven architectures using RabbitMQ, asynchronous messaging, queue processing, and micro-batching strategies to build scalable and resilient enterprise solutions.

In data engineering, I develop ETL/ELT pipelines and enterprise integrations using Databricks, PostgreSQL, Oracle, SQL Server, MongoDB, Redis, MySQL, and PostGIS, ensuring secure, high-performance, and reliable data architectures.

My experience includes integrating enterprise platforms such as SAP, ERP systems, CRM platforms, REST APIs, webhooks, cloud services, and third-party applications.

Earlier in my career, I worked with Computer Vision using OpenCV and developed data analysis solutions with Pandas.

Beyond implementation, I provide technical leadership by defining software architecture standards, mentoring engineering teams, reviewing system designs, and promoting best practices for scalability, maintainability, observability, and software quality.

Core Technologies: Python, FastAPI, Django, Flask, n8n, RabbitMQ, Databricks, OpenAI, Gemini, Vertex AI, PostgreSQL, Oracle, SQL Server, MongoDB, Redis, Docker, Git, REST APIs, Webhooks, RAG, AI Agents, Distributed Systems, Microservices, Enterprise Automation.

https://www.linkedin.com/in/paulo-lima-228a1a179/