How we work

How a normal week with us works.

How we start, what a normal week looks like, how live updates get shipped safely, and how your code and accounts stay yours the entire time. Same operating rhythm on every engagement.

Getting started in four stages

From signed contract to weekly live updates, in four stages.

Every project goes through these four stages. Each has a clear name, timing, and written outputs, so you know what to expect in week one and what a normal week looks like by week four. The timings assume we have access to a ready codebase.

First 48 hours

We pick up your code

We start within 48 hours of signing, once we have access to your code. We get access to your code, walk through your current setup, check your live environment, and send you a short first-impressions note. We tell you what looks healthy, what needs attention, and what we suggest working on first.

  • Access set up across your code repository
  • Walk-through of your current setup with screenshots
  • Short written first-impressions note
  • The first two weeks planned in the calendar
Week 1

Real work starts

We do not spend weeks ramping up. For a ready codebase, we pick the most useful thing we can deliver first and start work on it by Tuesday. The goal is to have that first update live the following week.

  • First work in progress by Tuesday (for ready codebases)
  • Record of decisions started
  • Communication channel agreed
  • Week one scope written down
Weeks 2 to 3

First live update, first report

For a ready codebase, your first live update reaches users within the first two weeks (once we have access to your code). Your first weekly update arrives on the Monday after your first full working week with us. It covers what went live, what is next, and the first decision we need from you.

  • First live update visible in your code history
  • Baseline set for release health
  • A tested way to safely undo a release
  • First weekly update delivered
Steady state

Weekly schedule

From week four onwards, every week follows the same simple loop. We build and release updates through the week. Every Monday morning you get a short written update. Monthly plan. 30 days' notice to cancel. Full refund if you cancel within 7 days of your first invoice.

  • A written update every Monday morning
  • Live updates released through the week, not late on Fridays
  • A short written review every month
  • Always ready to hand over if needed
A normal week

We build through the week and report on Monday.

Live updates go out during the week as work is ready. Every Monday morning you get a short written update with what happened last week, what is next, and any decisions we need from you. We avoid big releases late on a Friday so problems do not spill into your weekend.

  1. Monday

    Your weekly update arrives.

    Monday morning you get a short written update: what we did last week, what is next, anything that is blocking us, and any decisions we need from you.

    • Short written update on last week
    • This week's plan agreed
    • Record of decisions updated
    • One question flagged for you
  2. Tue to Thu

    We build, release, and decide.

    The middle of the week is for the actual work. Tasks move forward, changes land in your test environment, and live updates go out with a senior engineer signing off on anything important. If something is slipping, you hear about it mid-week, not only at the end.

    • Live updates released as work is ready
    • Test environment updated for your team
    • Any blockers surfaced in the open
    • Decisions written into the record
  3. Friday

    We close the week.

    Fridays are for wrapping up. We close off open tasks, draft Monday's update, and update the record of decisions. We avoid pushing big changes late on a Friday, so any problems do not spill into your weekend.

    • Open tasks closed off
    • Next Monday's update drafted
    • Record of decisions up to date
    • The team logs off on time
When something is delayed

If work slips, you hear about it mid-week, not only on Monday.

Not every week goes smoothly. Live systems break, third-party services fail, and sometimes a task turns out to be much bigger than it looked. When that happens, we handle it in writing, in the weekly update, and in your calendar. Four rules, the same every time.

  1. You hear mid-week

    The moment we know something is slipping, you know too. We do not wait until Monday to deliver bad news when we could tell you on a Wednesday.

  2. The next weekly update leads with what slipped

    The weekly update opens with what did not go live and why, in plain language. We do not hide a miss underneath the things that went well.

  3. A short written plan to catch up

    We send you a short note with the new target date, what we are trading off, and what we need from you. If we need your call, it is flagged at the top of the weekly update.

  4. Nothing falls off the plan without a decision

    Delayed work is rescheduled with an owner, dropped explicitly, or explained. Nothing disappears from the plan.

Who owns what

Everything we build and configure stays with you, in your accounts and repositories.

Every project is set up so you can hand it over at any time. If we stop working together tomorrow, you keep everything and can bring in anyone else to continue.

Code

Saved in your code repository from day one.

Every change is committed to your own GitHub or GitLab from the first day. We do not keep a private copy on our side.

Setup and accounts

All set up inside your own accounts.

Release setup, test environments, and live environments are configured inside your own cloud accounts and paid directly by you.

Documents

Kept in a place your team already uses.

The record of decisions, weekly updates, setup guides, and instructions all live where your team already reads them, not in a tool only we can see.

Keys and passwords

Reset when we leave.

Every key or password we used is handed back and reset the day we leave. Resetting them is part of the plan, not something you have to chase.

Talk to the founder

Talk to the founder directly.

Fifteen minutes on a video call. If we are not the right fit for what you need, I will say so. No sales pitch, no follow-up emails.

Rosaan, Founder of Antdragon
RosaanFounder · reviews every weekly update