>Remote-first backend work, from Urgench to clients I've never met in person

Most of my client work has happened without ever meeting the client in person — coordinating across time zones from Urgench, Uzbekistan, on systems that real businesses depend on daily. The tooling barely matters for this. What matters is a small set of habits that make asynchronous collaboration actually work.

Write the spec before writing the code

A short written spec — endpoints, request/response shapes, edge cases, what "done" looks like — forces the ambiguous parts of a task to surface before a day of code gets built on a wrong assumption. It also means a client in a different time zone can review and approve while I'm asleep, instead of a conversation stalling for eighteen hours waiting on a live call.

Small pull requests, reviewed asynchronously

A 40-file PR is a request for someone to block out an hour of focused attention — across a time zone gap, that hour might not arrive for a day. A focused, five-file PR with a clear description gets reviewed in the gaps between someone else's meetings. Scope discipline is a remote-collaboration skill as much as it's a code-quality one.

Demos over status updates

"It's going well" means nothing to a client who can't see the system. A two-minute screen recording of the actual feature working — or a staging link — replaces an entire status meeting and leaves no room for misaligned expectations about what "done" means.

Written English is the real interface

For international clients, the bottleneck was never Python — it was writing clearly enough in English that a spec, a bug report, or a PR description couldn't be misread. Investing time in writing precisely paid back more than any framework choice.

Where I work from

I'm based between Urgench and Tashkent, and open to relocating for the right in-office role in either city — but most of my production experience has been fully remote or hybrid, coordinating with distributed teams and clients abroad. Time zones and distance turned out to be a solvable logistics problem, not a technical one.

Takeaway

Ship weekly, measure in production, and over-invest in written communication. The stack is replaceable; the habits that make distributed collaboration trustworthy are not.