Founding a company in Sudan and running it from Dubai

Context

Four partners, in Sudan, Egypt and the UAE, with complementary pieces of a business and no company to put them in. I took the founder role: funding the incorporation from the UAE, setting the decisions, and building whatever the business needed in order to be taken seriously by an institutional client.

Standing a company up in a post-conflict economy is not a paperwork exercise. The requirements are circular — the bank wants a tax file, the tax office wants a lease and an operating licence, the licence wants a security clearance — and the currency moves fast enough that a cost recorded a week late is a cost recorded wrong.

What I did

• Drove the incorporation chain end to end and in the right order once I had mapped the dependencies: registration, security clearance, licence, tax file, and finally a corporate bank account — each one blocking the next, each handled by a different partner in a different country.
• Built the billing system the company invoices on. Sequential, never-repeating invoice and receipt numbers, branded PDF output carrying the company's electronic seal, and a register that is the single source of truth for what has been issued.
• Made every bill independently verifiable. Each one carries a QR code and a short verification code derived by HMAC from the bill number, which a client checks against a public lookup page on the company's own domain. A forged invoice does not verify, and the signing secret never leaves the local machine.
• Built the financial spine: a multi-sheet cost and payment tracker covering costs, payments, reserves, partner contributions and FX, so partners in three currencies are reconciling against one set of numbers rather than three recollections.
• Automated the reporting nobody had time to do by hand. A daily job reads the partner channel and the parallel-market FX channel, logs costs, payments and rates into the tracker, and transcribes Arabic voice notes locally so that decisions taken in a voice message end up in writing. Partners get a weekly financial report in English and Arabic.
• Positioned the company for institutional work — supplier registration with a UN agency, and pricing models built for tendered events rather than guessed at per job.

The problem underneath the invoice system

The trigger for building it was a client specifying the invoice format they would accept: company name, electronic stamp, electronic signature. The obvious response is a nicer template. The actual problem was different — in a market where a document is trusted on sight, a template is exactly what anyone can copy.

So the design goal became verification rather than presentation. The number is sequential and registered, the code is cryptographically derived rather than assigned, and the check happens on a page the client visits themselves. That converts trust in a supplier's paperwork into something a client can test in five seconds, which is the whole point when the supplier is new and the client is an institution.

What I would tell you in an interview

This is the clearest test I have had of whether I can run something without being able to walk over and look at it. Every control has to be explicit, every number has to have one home, and every decision taken in a voice note has to end up written down — otherwise the business runs on four people's memory, and memory disagrees. It is the same discipline as running a distributed operation inside a larger company, except that nobody hands you the templates.

Where this transfers

Multi-entity operations, distributed teams, and any business where documents have to be trusted by a counterparty who cannot verify you in person. The specifics here are Sudanese; the pattern — dependency-mapped setup, one canonical ledger, automated capture of decisions, and verification designed in rather than bolted on — is what I would bring to a new market entry or a franchise, partner or supplier network anywhere.