off somebody’s desk
The work here is other people’s routine: the month-end close nobody enjoys, the inbox that gets retyped into a CRM, the report somebody rebuilds every Sunday. If that sounds small, it is worth saying that it is most of what a business actually spends its hours on.
Before applying, read how we work. It is written plainly enough to put someone off, which is rather the point.
Open roles
Two seats, and one standing invitation
We are small and we hire slowly, so what is listed here is what is genuinely open. Both roles are delivery roles: you would be in front of clients within the first month.
01
OpenAutomation Engineer
- Location
- Muscat · hybrid
- Commitment
- Full-time
- Team
- Delivery
You build the things that stop happening. Most weeks that means reading a process nobody has written down, then turning it into something that runs unattended without anyone having to trust it blindly.
What you'd do
- Map a client's manual process end to end, then build the workflow that replaces it — integrations, rules, failure paths and all.
- Wire systems that were never meant to talk to each other: ERPs, CRMs, spreadsheets, inboxes, whatever the business actually runs on.
- Build assistants that draft and escalate rather than send blind, and prove they behave against real data before anyone depends on them.
- Stay after launch. Monitoring, tuning the rules and catching the edge case the process owner forgot to mention are the job, not an epilogue.
What we're looking for
- Two years or more building things that run without a person watching — automation platforms, integrations, scripts or backend services.
- Fluent enough in Python or TypeScript to be useful in week one, and unbothered by somebody else's API documentation.
- The instinct to ask what the work is for before asking what to build.
- English for the technical work and Arabic in the room, or a real intention to close that gap.
Useful, never required
- n8n, Make or Power Automate
- LLM tooling in production, not demos
- Odoo, Zoho or Dynamics
- Anything you have had to keep running
02
OpenERP & CRM Implementation Consultant
- Location
- Muscat · on site with clients
- Commitment
- Full-time
- Team
- Delivery
Half of this job is software. The other half is the finance manager who has closed the month the same way for eleven years and is not wrong to be careful about changing it.
What you'd do
- Run discovery: sit with the people doing the work, map what happens today, and find the hours actually worth going after.
- Configure and roll out ERP and CRM systems against that map rather than against a product tour.
- Move historic data across in stages, with the current way still running beside it, until the new one is proven on real work.
- Train the people who own the system after we leave, and write the runbook they will genuinely use.
What we're looking for
- Implementation experience end to end — Odoo, Zoho, Dynamics, Salesforce or similar — from discovery through to handover.
- Enough fluency in a P&L, a stock ledger or a sales pipeline to notice when a process is wrong, not just when it is slow.
- Plain writing. Most of this role is documents somebody else has to act on.
- Arabic and English, both used in front of clients.
Useful, never required
- Regional VAT and e-invoicing
- Data migration at volume
- Training and change management
- Send a CV
Open application
If neither role is yours and you would still be good here, send it anyway. We would rather read a CV early than have you wait for a posting that may not come.
Applying
How this actually goes
You send it
A CV and a few lines about something you built. No cover letter, and no form that asks you to retype the CV you just attached.
A conversation
Half an hour with whoever you would work beside, about work you have actually done. Not a quiz on the job description.
One piece of real work
A timeboxed exercise drawn from a problem we have had, and paid for at a day rate. Nobody works for free to prove they can work.
An answer either way
Within a week of the last conversation, with a reason attached. A no that arrives quickly is worth more than a maybe that never lands.