Hiring : AI Automation Engineer / n8n & AI Agent Developer

[Tomas_Maciulskas]

AI Automation Engineer / n8n & AI Agent Developer

Remote · Project-based cooperation · Freelance / Contractor / Long-term partnership

Synergy Effect is a 20+ year IT company now building a new strategic direction around AI agents, RPA and business process automation.

We work with real business processes: documents, emails, CRM / ERP systems, logistics, customer support, finance operations, internal workflows and other repetitive work that companies still handle manually.

We are looking for a practical AI Automation Engineer who can design and build production-ready automations using n8n, AI models, APIs, browser automation and modern agentic tools.

This is not a role for someone who only connects a few nodes and waits for perfect specifications. We are looking for someone who can understand a messy business process, ask the right questions, design the technical solution, estimate the work, build it, test it and deliver a reliable result.

What you will work on

You will build AI-powered workflow automations and agentic solutions for real business use cases, such as:

• document processing and OCR workflows;

• email and inbox automation;

• CRM / ERP integrations;

• invoice, order, logistics and accounting process automation;

• customer support and internal operations workflows;

• AI agents that classify, extract, validate, route and act on business data;

• browser automation for systems that do not have proper APIs;

• RAG and document search workflows;

• human-in-the-loop processes where AI needs validation or approval before taking action.

You will work with tools such as n8n, JavaScript, Python, REST APIs, OpenAI / Anthropic / Google / Azure models, OCR tools, databases, Google Workspace, Microsoft 365, ERP / CRM systems, Docker, Git and browser automation tools.

We are also interested in people who understand or have experimented with OpenClaw-style AI agents: autonomous or semi-autonomous agents that can work across apps, channels, files, browsers, scripts and tools — not just generate text.

What you will do

• Build production-ready automations and AI agents using n8n.

• Design workflow logic: data flows, branching, retries, fallbacks, error handling and human approval steps.

• Integrate business systems using REST APIs, webhooks, databases and third-party platforms.

• Write custom JavaScript inside n8n nodes for data transformation, validation and workflow logic.

• Use Python when needed for data processing, OCR structuring, RAG logic, scripts or backend automation.

• Work with LLM-based workflows: prompting, structured outputs, function calling, classification, information extraction and validation.

• Build browser automation when APIs are not available, using tools such as Puppeteer, Playwright, Selenium or similar.

• Test automations with real data, identify edge cases and make sure workflows can run reliably in production.

• Prepare estimates, technical tasks, documentation and clear status updates.

• Take ownership from business problem to deployed solution.

Required experience

You should have:

• practical experience with n8n;

• strong practical knowledge of JavaScript;

• good understanding of REST APIs, webhooks, JSON and authentication;

• basic SQL knowledge and understanding of data structures;

• practical understanding of how AI agents and LLM workflows work;

• ability to model business processes and turn them into workflow logic;

• ability to work independently in a remote, asynchronous environment;

• clear written communication and good English;

• ability to estimate work, communicate risks and take responsibility for delivery.

Strong advantages

Experience with any of the following would be a strong plus:

• OpenClaw or similar autonomous AI agent frameworks;

• building agent workflows that can use tools, files, browsers, scripts or external systems;

• creating or using agent skills, plugins or reusable automation modules;

• running AI agents locally or in controlled / sandboxed environments;

• Python for data processing, OCR, RAG or backend automation;

• OCR tools: Google Vision, AWS Textract, Azure Document Intelligence, Tesseract or similar;

• browser automation: Puppeteer, Playwright, Selenium or Apify;

• Google Workspace API or Microsoft 365 / Outlook / SharePoint API;

• ERP, CRM, accounting, logistics or e-commerce system integrations;

• RAG, embeddings, vector databases and document search;

• Docker, Git, cloud infrastructure or self-hosted automation environments;

• AI-assisted development tools such as Claude Code, Cursor, GitHub Copilot or similar;

• production systems, not only demos or proof-of-concepts.

Technologies you may work with

• n8n

• JavaScript

• Python

• SQL

• REST APIs / Webhooks

• OpenAI / Anthropic / Google / Azure AI models

• OpenClaw or similar AI agent frameworks

• OCR tools

• Google Workspace API

• Microsoft 365 / Outlook / SharePoint API

• PostgreSQL / MySQL / BigQuery / Google Sheets

• Git

• Docker

• Puppeteer / Playwright / Selenium / Apify

• CRM / ERP / accounting system APIs

• RAG, embeddings and vector databases

• Claude Code / Cursor / GitHub Copilot or similar tools

Who we are looking for

We are looking for someone who enjoys understanding how businesses actually work.

You should be able to take a real-world process that is unclear, repetitive or messy and turn it into a structured automation that works reliably.

You understand that an AI agent is not just a prompt. It is a combination of process logic, data, integrations, permissions, validations, error handling, memory, monitoring and stable execution.

You are comfortable asking questions, challenging unclear requirements and proposing practical solutions. You do not need every detail handed to you before you can start thinking.

This role is best suited for a mid-to-senior automation specialist who has already built real workflows and can take ownership of delivery.

Who is not a fit

This role is probably not for you if:

• you only connect basic automation nodes without understanding the business process;

• you have only worked with Make, Zapier or Power Automate and have no real n8n experience;

• you wait for fully prepared specifications before doing anything;

• you are interested only in a strictly defined full-time role;

• you build demos but do not think about testing, edge cases, security and production reliability;

• you believe AI will solve everything by itself.

AI does not replace architecture. A model without process logic, data structure and validation is just an expensive guessing machine.

Cooperation model

This is a remote, project-based cooperation opportunity.

At this stage, we are not necessarily looking for a full-time employee. We are looking for a flexible specialist, freelancer or contractor who can take ownership of specific AI automation projects and deliver them end-to-end.

The workload may vary depending on active client projects. If cooperation works well, it may grow into:

• a broader long-term partnership;

• a larger project load;

• a permanent role;

• or eventually an AI Solutions Architect / Lead AI Engineer position.

Compensation will be discussed individually based on experience, project scope and expected workload.

What we offer

• Work on real AI agent and automation projects, not theoretical demos.

• A chance to help build Synergy Effect’s new AI Agents & RPA direction.

• Diverse business cases across logistics, finance, e-commerce, customer support, document management, ERP / CRM automation and internal operations.

• Room to propose solutions, not just execute tasks.

• Remote and flexible project-based cooperation.

• Opportunity to grow into a longer-term partnership or leadership role.

• Practical, business-driven work where automation must create measurable value.

Recruitment process

1. Short introductory call about your experience and motivation.

2. Technical discussion based on a real automation or AI agent scenario.

3. Practical mini-task or solution design exercise.

4. Final discussion about cooperation model, availability and compensation.

How to apply

Send your CV or LinkedIn profile to info@s-e.lt and briefly answer:

1. What n8n, AI agent or RPA solution have you built?

2. Which systems did you integrate?

3. What was the business problem?

4. What solution did you create?

5. What was the result?

6. Have you worked with OpenClaw or similar agent frameworks?

7. What is your availability for project-based remote cooperation?

8. What is your preferred cooperation model and compensation expectation?

If you have a workflow diagram, GitHub repository, demo, case study or technical write-up — please include it.

Important: practical n8n experience is mandatory for this role. Experience with Make, Zapier or Power Automate can be useful, but it does not replace real n8n experience.

[Pathfinder_Automate]

Hello @Tomas_Maciulskas , welcome to n8n community, I have worked and have experience with n8n and l will love to collaborate with you on this you can schedule a call Here and you can checkout my upwork profile Here, for my pastworks and certifications

[moosa]

Hi Tomas, welcome to the n8n Community, I’ve sent you an email with my experience, workflow references, and profiles. I believe my background in n8n, AI automation, OCR systems, and API integrations aligns very well with this role. Looking forward to connecting with you.

[Aryan_Pmedia]

Just sent you an email with all the details and I can get it done for you.
Looking forward to speak soon!

[Mihail_Rogal]

Hi! Replied via gmail

[Hamza426]

Hi Tomas,

I read the post twice. Not for the requirements list — for what’s behind it.

A 20-year IT company building a new AI direction needs one thing above everything else: someone who doesn’t need to be managed. Someone who can walk into a messy, undocumented business process, figure out what actually needs to happen, and deliver something that works in production — not something that works in a demo.

That’s a harder profile to find than the job post makes it sound. Most people applying know n8n. Far fewer can take ownership of the full arc from business problem to deployed solution without constant hand-holding.

Here is where I stand on each of your requirements.

On autonomous agents and real production systems

I run a customized version of OpenClaw — the same framework you referenced in the post. I didn’t just install it. I rebuilt significant portions of it and extended it with 22 custom scripts covering Gmail SMTP/IMAP operations, a Google Maps scraper that generates up to 2,200 qualified leads per run, a headless Playwright browser bot for social media operations with both burner and real browser profiles, a Chrome extension that gives the agent browser visibility, and an autonomous execution loop that runs on cron without requiring a human to press start.

The memory architecture I built uses 11 structured markdown brain files. The agent reads its own state at startup, picks the highest-priority task, executes it, verifies the result against real retrieved data before logging it as complete, and updates its memory before sleeping. It cannot mark a task done based on assumption. That was a deliberate architectural choice — not a default feature of the framework.

The agent has run 150+ cold outreach emails across 5 separate campaigns covering different industries: SMMA leads, Denver roofers, Spokane HVAC, Med-Spa buyers. Lead fetching, email writing, sending, reply monitoring — all without manual execution per email. That is not a demo. It is running on my machine right now.

I also integrated Ollama for local AI inference so the agent can make decisions and write emails entirely on-device when needed, without external API dependency.

On the specific work Synergy Effect needs to do

You listed document processing, email automation, CRM/ERP integrations, invoice and logistics workflows, customer support operations, and browser automation for systems without proper APIs. I want to address these directly.

Document processing and structured data extraction: I work with LLMs for classification, information extraction, and structured output generation — not just text generation. The model is one validation layer in a larger process, not the entire solution.

Email operations: I have built and run SMTP/IMAP pipelines with App password authentication, SSL handling, reply fetching, and bounce monitoring. Not theoretical — these have sent over 150 emails in live campaigns.

Browser automation: The social-bot.py script in my setup uses headless Playwright with session management, profile handling, and a browser gateway that stays stable across restarts. I also built pcctl, a command-line browser controller for agent-directed browsing.

CRM and data routing: I have built lead enrichment pipelines connecting scraping, enrichment logic, qualification scoring, and Airtable CRM writes — with outreach triggers firing automatically when leads cross a confidence threshold.

Human-in-the-loop escalation: Built into the architecture. The agent flags tasks that exceed its confidence boundary rather than guessing and logging a false completion.

On n8n specifically

I have built n8n workflows covering webhook-triggered pipelines with conditional branching, API integrations connecting CRMs and Google Sheets, lead enrichment flows with validation and Airtable writes, and LLM-integrated workflows using structured outputs for classification. I understand error catch nodes, retry logic, fallback paths, and silent failure prevention. I have not shipped a paid client n8n project yet — my production automation work has lived in Python and the OpenClaw stack. I am telling you this directly because you will find out on a technical call anyway, and I would rather you have the accurate picture now.

What I can tell you is that the underlying capability transfers directly. Anyone who has designed an autonomous agent with error handling, memory, validation, and execution verification understands workflow logic at a level that goes deeper than connecting nodes.

On working style

You wrote that you are not looking for someone who waits for fully prepared specifications before doing anything. That matches how I work. I am used to walking into unclear processes, mapping them myself, identifying what is actually the problem versus what the person thinks is the problem, and proposing a structured solution before building anything. I work async, document decisions, communicate blockers early, and do not need daily check-ins to make progress.

Availability and cooperation model

Available immediately for project-based remote cooperation. I can align with EU/UK time zones. I prefer fixed-price agreements on scoped projects — it means I am accountable to delivery, not hours. Rates depend on scope and complexity and I am happy to discuss that on the intro call.

If it would move things forward faster, I am ready for your mini-task or technical scenario exercise now rather than waiting for a call. That is probably the most efficient way for both of us to assess fit.

Hamza
Founder, AXONA

itsameerhamza203@gmail.com

[kevork_lepedjian]

@Tomas_Maciulskas

A

Hi, I’m very interested I’ve been building real n8n + AI agent workflows involving OpenAI integrations, browser automation, APIs, RAG pipelines, CRM/workflow automation, and custom JavaScript logic, and this kind of practical business-focused automation work is exactly what I’m looking for

[soufiane_business]

Hi Tomas,

This is a strong fit for me because the work is not just node wiring. I build n8n and AI automation around real process logic: inputs, validation, routing, retries, human approval, logs, and handoff docs.

Relevant fit:

• n8n workflows with JavaScript Code nodes, APIs, webhooks, databases, Google Sheets, CRM-style flows, and error handling
• OpenAI/Claude workflows with structured output, classification, extraction, and decision steps
• browser automation when an API is missing, using Playwright-style flows where appropriate
• practical production habits: test data first, no silent failures, clear scope, and simple docs

Best first step: give me one messy but valuable process and I will scope it into a paid first milestone. For example: document/OCR intake, inbox triage, CRM routing, invoice/order flow, or browser automation around a system without an API.

I prefer a small paid trial task first so the result is easy to judge. I can start async, ask only the questions needed, and return a working first loop plus test notes.

[Basit_Adefola]

Hi,

I’m interested in the AI Automation Engineer / n8n & AI Agent Developer role.

I build automation workflows using n8n, Make.com, Zapier, and the OpenAI API — including multi-step AI agents, CRM integrations, webhook pipelines, and business process automation. I’m fully remote and available immediately for project-based or long-term collaboration.

Happy to share examples of workflows I’ve built. Looking forward to hearing from you.

Adedoyin Basit Adefola
LinkedIn: linkedin.com/in/adedoyin-adefola-b60a85405

[syed_noor]

Tomas — interested. The “messy process → structured solution” framing in your post is the right filter, and the explicit “not a fit” list is the cleanest screening criteria I’ve seen for this kind of role.

Different agent stack than @Hamza426 (who set a high bar above): I run Claude Code + n8n-mcp as the build loop, which gives the autonomous architecture/iteration velocity you’re describing without OpenClaw specifically. Background is production-systems engineering (multi-process trading infrastructure, deterministic + LLM-bounded designs, audit logs from day one) — applying that discipline to n8n builds the last few weeks.

I’m emailing the full application (answers to your 8 questions) to info@s-e.lt today. Happy to do the practical mini-task as the next step after the intro call.

[nxning_ai]

Hi syed_noor!

Saw your post about Hiring : AI Automation Engineer / n8n & AI Agent Developer.

I specialize in n8n workflow automation and have built end-to-end systems for:

• CRM pipelines & lead generation automation
• Multi-platform integrations (Google, Salesforce, HubSpot, custom APIs)
• Custom AI agents with n8n (OpenAI, Anthropic)
• Production error handling and monitoring

Happy to discuss your requirements. Feel free to DM or email!

[jettylee]

Hi Tomas / Synergy Effect team,

I am interested in this AI Automation Engineer / n8n & AI Agent Developer role.

My background:

• 12+ years full-stack engineering and technical leadership
• Current focus: AI agents, LLM workflow automation, API integrations, Web3/RWA systems, and operational tooling
• Strong backend/integration experience across Node.js, Python, Java, databases, REST APIs, webhooks, Docker, Git, browser automation, and production system design
• Experience designing AI-assisted workflows with clear inputs, outputs, logs, exception handling, and human approval gates

Relevant workflow pattern I can build:

1. Intake from email, forms, CRM, documents, or internal tools.
2. Classification and data extraction using deterministic rules first, then LLM reasoning where useful.
3. API/webhook integrations with Google Workspace, Microsoft 365, CRM/ERP, or custom systems.
4. Human approval gates for sensitive or client-facing actions.
5. Logs, retries, and exception queues so the automation can be trusted in production.

I am based in Chiang Mai, Thailand (UTC+7), available for project-based remote cooperation, and comfortable working asynchronously with EU overlap.

Relevant profile: https://www.jettylee.com/

Best,
Jetty Lee

[Harrison_Pettigrew]

Interested — sending a full application to info@s-e.lt, but wanted to add a bit of context here first since the post is detailed enough to warrant it.

The “who is not a fit” section describes exactly how I think about this work too. An AI agent isn’t a prompt with a trigger — it’s process logic, data structure, validation, error handling, and stable execution. I’ve seen enough automation projects fail because someone treated the LLM as the architecture that I default to designing the workflow logic first and fitting the AI layer in where it genuinely reduces manual decision-making.

On the specific stack you mentioned:

• n8n + JavaScript: daily driver, custom code nodes for data transformation and validation logic, sub-workflow architecture for reusable components and maintainable error handling
• LLM workflows: structured output extraction, classification pipelines, function calling — and yes, knowing when not to use a model because a deterministic check is faster and cheaper
• Browser automation: Playwright for systems without proper APIs, used it for CRM data pulls and order management systems that predated API access
• OCR / document processing: worked with Google Vision and Textract for invoice and document intake workflows feeding into CRM/ERP
• OpenClaw-style agents: haven’t used OpenClaw specifically but have built semi-autonomous agents in n8n with tool-use patterns, memory nodes, and human-in-the-loop approval gates — happy to discuss what you’re building there in the technical discussion

The business cases you listed — logistics, finance, document processing, CRM/ERP integrations — are the areas where I’ve done the most production work. Not demos.

Will send the full breakdown with examples to info@s-e.lt.

[Harrison_Pettigrew]

Hi @Tomas_Maciulskas — the Synergy Effect post is well-written and the “who is not a fit” list is genuinely useful. Let me respond to that framing directly.

The part that stands out to me: a 20-year IT company building a new AI direction means the processes you’re automating already exist and have edge cases baked in through years of use. That’s different from a startup building workflows for a greenfield process. It means I’d expect undocumented exceptions, legacy system constraints, and internal logic that nobody wrote down. That’s the kind of environment I find more interesting to work in, not less.

On your specific requirements:

n8n: I build complex multi-step workflows in production — sub-workflows for reusability, error handling with fallback branches and alert paths, webhook-triggered flows, and modular architecture so individual pieces can be updated without touching the whole system. JavaScript Code nodes for any data transformation or business logic that the built-in nodes don’t cover cleanly.

LLM integration: OpenAI and Claude APIs for document parsing, email classification, structured data extraction, and decision routing. I focus on getting reliable JSON outputs from LLM calls rather than hoping the model formats things correctly — structured output prompts, validation steps, and fallback handling when the model produces something unexpected.

Browser automation: Playwright for scraping or interacting with systems that don’t have APIs. Useful for the kind of legacy internal tools a 20-year-old IT company is likely running.

Production habits: I test on realistic data before considering anything production-ready. Documentation comes with the build, not as an afterthought.

Background: manufacturing and operations before going deep on automation — which means I think about what happens when a workflow fails at 2am, not just what happens in the happy path demo.

I’m emailing to info@s-e.lt with answers to your 8 questions. Quick question first though: are most of the client processes you’re automating document-heavy (PDFs, forms, invoices) or more workflow coordination and system integration work?

[timai11]

This is close to the kind of automation work I’m actively focused on — especially the OpenClaw-style agent part.

I’m not going to overclaim senior enterprise n8n history, but I can be useful for small practical slices: messy input → structured extraction/classification → Sheet/CRM/database handoff → human approval → logs/failure reporting.

A good test task for me would be one workflow map or small build around inbox/docs/CRM automation, with clear edge cases and handoff notes. I’ve also emailed info@s-e.lt with a more complete intro.

[hermesautomation]

Hi Tomas,

I’m interested in a small project-based pilot rather than starting with a big undefined scope.

Based on your post, I’d suggest starting with one practical workflow:

1. Input: documents / emails / CRM records

2. Processing: extract key fields, classify urgency/type, detect missing data

3. Output: CRM/ERP update + notification + audit log

4. Reliability: retry, error log, and a short handover note

I’ve prepared a small public demo pack showing the type of work I can deliver:

GitHub - grainnecortez72708-png/income-automation-demo-pack: Small automation demos: web monitoring, CSV...

It includes website change monitoring, CSV/lead cleanup, and lead/message classification.

If useful, I can do a first paid pilot with a narrow scope and deliver a working version quickly.

Payment can be project-based; USDT is also fine if that is easier.

[Suhail_Narot]

Hi Tomas — I’ve already sent an application to info@s-e.lt, but wanted to reply here too since this description maps very closely to what I do.

I’m Suhail Narot, an AI automation engineer running Fajr AI. The framing of this role resonates: ‘someone who can understand a messy business process, ask the right questions, design the technical solution, estimate the work, build it, test it and deliver a reliable result.’ That is how I operate.

What I’ve built in production:

- AI booking agent for a local services business: inbound WhatsApp → Claude intent routing → Google Calendar availability check → booking confirmation and reminder sequences. Multi-channel (FB + IG DMs also wired via Meta Graph API).

- Telegram → Zoho Invoice automation for field reps generating invoices on-site with validation rules and audit logging.

- Apollo + LinkedIn lead qualification pipeline feeding personalised outreach sequences.

- Daily AI content pipeline: Claude-scripted video → HeyGen render → scheduled Meta post.

For the document/email/CRM work Synergy Effect handles: I’ve integrated REST APIs for SMTP transactional email, Zoho, WhatsApp Business (WATI), and webhook-driven ERP-style systems. I’m comfortable reading messy processes, proposing a scoped v1, and iterating from working software rather than specs.

Available immediately, remote, UTC+2.

-– Suhail Narot | Fajr AI | narotsuhail@gmail.com

[maxim_makselyanov]

Hi Tomas. This is close to the kind of production automation work I prefer: messy business process → clear workflow → tested delivery.

I work across n8n, backend/API integrations, LLM workflows, browser automation, CRM/admin systems, and human-approved automation where reliability matters.

I would start with one paid pilot around document/email intake, structured extraction, CRM/ERP update, human approval, and error logging. That gives you a real signal on delivery before a broader partnership.

Happy to discuss one concrete workflow first.