Support and further development
Changes to a system that already works, and already gets in the way
Support is not an unlimited chat of “please also fix this”. It is a queue: bugs, speed, new functions, and careful development of the architecture. I join existing sites, accounts, and internal systems, when the code can be understood and a change does not become a lottery.
What it takes on
- 01
Fix a bug
Reproduce it, fix the cause, and leave a trace so it does not come back in silence.
- 02
Make it faster
Find what slows the page load or the employee's work, and remove that specific cause.
- 03
Add a function
Fit a new scenario into the current architecture, instead of attaching it at the side.
- 04
Prepare for growth
Separate roles, data, and deploy, so the next batch does not start with an excavation.
Who it is for
Teams without their own developer
The product is live, and there is no one who can change it safely.
Projects after another team
You need to understand what the code is doing before new modules can be promised.
Companies with regular changes
Every month there is a small queue, and finding a new contractor each time costs more.
If the project cannot be built, there is no access, or every touch breaks a neighboring module, it is more honest to stop and talk about recovery. Quiet support is not a promise in that case.
What can be built
Bugs
Reproduction, a fix, and a check of the neighboring scenarios.
Speed
Page load, and the heavy places in the interface.
New functions
A queue of changes with a clear boundary.
Scaling
Data, permissions, and a build that can take the next step.
Ongoing support
A regular slot, not random edits at night.
How it looks in practice
- 01
The site sometimes fails to send a request
First we catch the condition, then we fix it. At the same time we check that letters are not lost in silence. “It did not happen again for me” is not a diagnosis.
- 02
A new role is needed in the account
We look at how the current permissions are built. If they are wired into the screens, the rule is pulled out carefully first. Then the role is added.
- 03
The product is live, and the developer has left
A short audit: how to run it, where the secrets are, and what is dangerous to touch. After that, the queue can be taken on. Without the audit, support is guesswork.
How development goes
- 01
Access and a run
The project builds, the environment can be repeated, and secrets are not in the chat.
- 02
A short review
What hurts, what is dangerous, and what can be done at once.
- 03
The queue
Tasks with a boundary, not a stream of messages.
- 04
The change
A fix, a check of the scenario, and a note.
- 05
Release
It is clear which version is on the server.
- 06
The next step
What remains, and whether it should be a project of its own.
Technologies
- TypeScript
- React
- Next.js
- Node.js
- PostgreSQL
- The project's existing stack
For each task you receive a change in the repository, a short note of what was checked, and a note of what was left alone on purpose.
Demonstrations on this topic
There is no public demonstration in this direction yet. Neighbouring interfaces are in the laboratory.
How to commission it
A separate scope
We take one mechanic, one link between systems, or one set of screens that can be finished and checked.
Inside a larger product
The interface, the game, or the automation sits inside a larger project. It does not live apart from the data.
Support
After handover there is a queue: fixes, new scenarios, and someone watching the system.
Questions
Do you take on code written by someone else?
Yes, after a short look. If the project cannot be started, or a change to any line is unpredictable, I say so before a retainer, not at the end of the month.
How is support different from a new project?
Support changes an existing system in small finished tasks. A new module, with its own logic, roles, and data, is more honest as a project with its own boundary.
Can fixes be done from messages in a messenger?
For an urgent bug, yes. A queue needs a list: what to do, and how to tell that it is done. Otherwise the month ends, and the work has no boundary.
Do you guarantee that nothing will break?
No. That does not happen in a live system. I check the scenario that was touched, and the neighboring paths that depend on it. A claim of zero risk usually means the risk was left unnamed.
How is access handed over?
Through separate channels, not in one letter together with the password to everything. The secrets stay with you. They do not enter the repository.