Consulting / How to start
Tell me what the system does and where it goes wrong. Within a week I send back a written report: what is broken, what I would fix first, and how long that would take. The review costs 500 EUR. If you hire me for the work afterwards, I take the 500 EUR off what the work costs.
Getting a model to draft well is usually the easy part. The problem shows up the day it drafts something wrong and nobody notices until a customer does. So I build the approval step and the log of what the model did at the start, rather than adding them after something has gone out.
Someone spends part of every week collecting the same data and writing the same summary. When that person is busy, the report does not get written at all. I turn the job into a tool you run, and keep the output plain enough that you can check the answer yourself.
Arguing with a plan is cheap while it is still a document. Once it is code, the same objection means a rewrite, so people stop raising it. This is the smallest job I take. Often it is the only one someone needs.
Deciding what a product should look like and building it from someone else's decision are two different jobs. I do the second one. Send a design that already exists and I will build the interface from it.
I hand over the code and the notes needed to run it, written so that someone else can. I do not stay on a monthly fee to keep it running, because that turns a finished job into a dependency on one person.
Joining a team and working down its task list for a few months is real work, and often the right hire, but it is not what I do. I take on a problem whose end you can describe before it starts.
One email describing what you have and where it is going wrong is enough to start.
