Mujammil
Taking new workStart a build

How I work

Five stages. You can
see it running from
the first week.

Nothing here is unusual. It is written down because "we'll keep you updated" is what everyone says, and a deployed URL in week one is the only version of that promise you can actually check.

  1. 01

    One call, no deck

    45 minutes

    I want to hear the thing that goes wrong every week, not the feature list. By the end of the call I can usually tell you whether the project is worth building at all, including when the honest answer is a spreadsheet and a rule, not software.

    You end up withA written summary of what I heard, and what I think it costs

  2. 02

    Scope, fixed, in writing

    2–4 days

    What is in, what is explicitly out, and the assumptions the price depends on. The out-list is the important half. Most projects go wrong because nobody wrote down what was never included.

    You end up withFixed price, dates, and an assumptions list you can argue with

  3. 03

    Build, in the open

    2–10 weeks

    A deployed preview URL from the first week. Not screenshots, not a demo on my machine. A link you can open on your phone in a meeting. You see progress continuously, so nothing at the end is a surprise.

    You end up withA live environment, updated as I go, and a weekly written note

  4. 04

    Real data, real users

    1–2 weeks

    The stage most projects skip. Software behaves differently when the table has 40,000 rows and someone is using it on a bad connection in a car park. This is where I find the things a demo never shows.

    You end up withLoad-tested, on real data, with the slow queries already fixed

  5. 05

    Handover, then support

    Ongoing

    Documentation written for whoever maintains it next, which is usually not me: architecture, runbook, and the decisions that will otherwise look arbitrary in a year. Then a defined support window and a rate for after it.

    You end up withDocs, a runbook, and someone who answers when it breaks

Rules I do not break

Four commitments.

  • You get a URL, not a screenshot

    From week one. If I cannot show you something running, I have not made progress worth reporting.

  • Bad news travels fastest

    The day something slips, you hear it, with a revised date attached. Late news is the only kind that actually costs you a client.

  • I write down what is not included

    Every quote carries an explicit out-list. Almost every dispute in this trade is about something nobody wrote down.

  • The last invoice is worth coming back for

    Final payment is never a token amount. It has to be large enough that finishing properly is in my interest.