Careers

Build and ship the workflows that replace manual, repeat processes for clients abroad.

• Map a client process, then build it in n8n, Make or Zapier
• Wire integrations to CRM, ERP, help desk and email APIs
• Add retries, fallback paths, alerts and an exception queue
• Test on 20–50 real cases and report before-and-after numbers

You likely have: working experience with at least one automation platform, comfort with REST APIs and JSON, and the patience to debug someone else’s data.

Build agents that plan their own steps and call tools, inside limits that keep them safe to run.

• Design tool interfaces, permissions and spend caps for agents
• Implement action logging and rollback paths
• Write evaluation sets and measure accuracy before release
• Decide, honestly, when a plain workflow would be better

You likely have: Python, experience calling LLM APIs in production, and an instinct for what can go wrong when software acts on its own.

Turn scattered prompts into a versioned system with test cases that prove each one still works.

• Write and version prompts against a defined output contract
• Build gold sets and regression tests for prompt changes
• Tune for cost and latency, not only for quality
• Document why a prompt is written the way it is

You likely have: strong written English, a methodical testing habit, and real examples of prompts you improved with measured results.

Sit between the model and the client’s customer, and catch what the model gets wrong.

• Review AI output against a written guideline and a gold set
• Approve, edit or reject, and log the reason every time
• Feed recurring errors back to the engineers
• Help write the guideline when a new queue starts

You likely have: careful reading, consistency under volume, and comfort being measured on agreement scores. Domain background in payroll, finance or HR is a plus.

Clean, structure and connect the data that every automation and agent depends on.

• Build pipelines from client systems into usable, structured data
• Deduplicate, validate and correct records with a documented method
• Prepare evaluation and training-free reference data sets
• Keep data access least-privilege and auditable

You likely have: SQL, Python, experience with messy real-world records, and the discipline to document assumptions rather than silently fix things.

Run content, campaigns and demand generation with AI in the pipeline and a human editor on the output.

• Build draft, edit and fact-check pipelines for content at volume
• Own brand voice specifications the model and editors both follow
• Run email segmentation, campaign reporting and ad automation rules
• Measure what actually produced enquiries, not impressions

You likely have: strong English writing, hands-on experience with an AI writing workflow, and the honesty to kill content that is not working.