How the work is staffed

A founder-led shop, and the specialists we bring in by name

There is no team wall on this page, because a grid of headshots is the cheapest signal on the internet to manufacture. What we can show instead is how the work is actually staffed: who does it, when a specialist joins, what we commit to in writing, and what happens on the projects where we are simply the wrong size.

Book a call See how an engagement runs →

Two colleagues talking by an office window, one holding an open laptop

1: accountable name per engagement, from baseline to handoverNamed in the proposal, with a substitute named beside them and the schedule consequence stated in the same paragraph.
0: account managers between you and the person writing the codeA structural fact about our size, not a service promise. It stops being true above a certain headcount, which is one reason we are in no hurry to grow.
4: systems in production built by the same people you would be hiringThe car marketplace, beron.mon.bg for the Ministry of Education, the email and Instagram NLP module, and the beauty-brand email automation.
1 of 4: of those systems you can open without asking us for anythingberon.mon.bg is published by the ministry and loads in your browser. For the other three the reference is a conversation, and we make the introduction before you sign.

Who does the work

The person you meet on the first call is the person who writes the code and signs the handover. That is the entire structural claim, and it is one a fifty-person consultancy cannot make: there is no discovery team passing a document to a delivery team that never sat in the interviews.

Practically, it means one accountable name runs from the baseline measurement through to the recovery test at handover. When something breaks at 19:00 on a Thursday, you are not routed to a partner who has to go and ask somebody else what your system does.

It also means being honest about hours. A founder-led shop that claims unlimited availability is either lying or about to be. We will tell any client how many engagements are running concurrently, and we would rather lose a project to the calendar than staff it badly.

How specialists join a project

Beyond the core work we contract specialists per engagement, and each one is named in the proposal along with the scope of their part and the days budgeted. They are not listed on this page as staff, because they are not staff — a specialist network presented as headcount is exactly the small dishonesty this page exists to avoid.

The tradeoff is real, so here it is. Contracted specialists cost more per day than an in-house equivalent and they are not available at zero notice. What you get in exchange is people who do only that one thing, and no bench being paid for out of your project.

Data engineering — warehousing, entity resolution, pipeline reliability when a system of record is in scope
Interface design — where an operator or a customer actually touches the system
Bulgarian-language linguistics — lexical, terminology and content work, the discipline behind the Ministry dictionary
Legal and data-protection review — a DPO or counsel signs what an engineer should not be signing

What you can hold us to

Since we cannot offer a wall of logos or a certification badge, we offer commitments that are checkable while the work is happening rather than afterwards. They go into the proposal, not into a values page, and all of them are observable inside the first month.

Two of them are about us rather than about the system. We name a substitute for the accountable person and state the schedule consequence if that person is unavailable. And we write down the condition under which we would recommend stopping — the number that would make phase two the wrong purchase.

A fixed fee on the first phase, with the deliverable named rather than described
An acceptance threshold written and countersigned before the first result is seen
Your repository, your cloud account, your credentials — from the first commit, not at handover
A handover document validated by your engineer running a full recovery from it while we watch and say nothing
A named substitute, and the stated schedule consequence if the accountable person is unavailable
A written stopping condition: the result that would make us recommend against phase two

When we are the wrong size

There is a shape of project we cannot take: parallel workstreams across several countries, a dozen people on site, a programme that needs a project office before it needs an engineer. We say so on the first call and name firms who do that work, rather than accepting the brief and going recruiting.

There is also a shape we can take but should not lead — a build whose binding constraint is a change to how a department operates rather than the software. We can do the measurement and the integration; the organisational half needs someone inside your business with authority, and if that person does not exist the project becomes a tooling purchase with a governance layer on top.

Saying this costs us work. It costs considerably less than the alternative, which is being three months into a programme that was mis-sized in the first meeting, opposite a supplier who cannot say so out loud because the invoice depends on it continuing. The size question is cheap to answer in the first hour and expensive to answer in the second quarter.

Why there are no photographs on this page

A headshot grid signals scale, which is precisely the signal we would be borrowing rather than earning. It also ages badly at this size, where a page of five faces has to be quietly edited every time an arrangement changes, and quiet edits are how a site starts disagreeing with itself.

The things worth verifying about a small supplier are not faces. They are which systems are actually live, what the contract says, and whether the person on the call can answer a technical question without leaving the room. One of our four systems is open in a browser right now; the other three we will discuss in detail with anyone serious.

The size question is cheap to answer in the first hour and expensive to answer in the second quarter.

Which is why we raise it on the first call, before scope or price.

When we are the wrong choice

If your procurement requires a supplier above a certain employee count, we will not clear the gate, and we would rather you knew that in week one than in week nine.
If you need people in your building every day, we work remotely with periodic on-site sessions. That is a working model, not an opening negotiating position.
If the engagement depends on us assembling a team of six inside a month, we cannot do it. Our network is deep in a few disciplines rather than wide across many.
If you want one supplier to own strategy, build, change management and the training programme, we are the wrong shape. We do the measurement and the build, and we will say who should do the rest.

Questions we get asked

Who will actually be on our project?

The accountable person is named in the proposal, along with any specialist, the scope of their part and the days budgeted for it. There is no bait-and-switch structure here where a senior name appears in the pitch and a junior appears on the calls, because there is nobody junior to substitute in. If the composition has to change mid-engagement, you hear it before it happens and you can stop the phase.

What happens if the person who knows our system leaves or gets ill?

This is the honest risk of buying from a small firm, and we will not talk around it. Three things reduce it: your code and infrastructure definitions live in your repository from the first commit, the handover document is written during the build and tested by your own engineer, and every proposal names a substitute with the schedule consequence stated. What none of that removes is delay — a serious absence will move dates, and we would rather write that here than discover it with you.

How many clients do you work with at once?

A small single digit, and we will tell you the current number on the first call. We turn work down when the calendar is full rather than accepting it and then finding people, because the version of this business that hires to fit demand is the version where you stop talking to the person who writes your code.

Do you have certifications — ISO, partner tiers, SOC 2?

No, and we are not going to display badges we have not earned. What we can put in a contract instead is specific: where data is processed, whether a vendor trains on inputs, who holds which credentials, retention periods, and a DPA. If certification is a hard gate in your procurement, we will not pass it, and that is a reasonable rule for you to have.

Can you work alongside our internal engineering team?

Yes, and it is often the better arrangement — we do the measurement, the integration and the parts nobody internally has time for, while your team keeps the knowledge. In that mode we work in your repository, on your review process, with your engineers on the pull requests. The handover test still applies: at the end, one of your people runs a full recovery from the runbook unaided.

Keep reading

How we work, phase by phase Four systems, and what each taught us Open roles and how we hire

Ask who would be on it before you ask what it costs

Bring the project. In thirty minutes we will tell you who would be accountable, which specialist the work needs, roughly how many days each part takes, and whether this is a shape we should be leading at all.

Book a call See our projects →

If the answer is that you need a bigger supplier, you will hear it on that call, and we will name two.