01About
A studio built around the part everyone skips.
02Why Lluyot Labs exists
Most models die in a notebook.
A model is written, validated, presented — and then nothing happens. Not because it was wrong, but because the distance from a correct result to a product someone can act on turns out to be enormous, and nobody in the room owns that distance.
Lluyot Labs was founded to own it. The studio was set up so that the same people who formulate a problem also build the interface that exposes it, choose the words that describe its uncertainty, and answer for what happens when a user relies on it. That arrangement is unusual, and it is the reason the results hold together.
It also imposes a limit we accept: we work on fewer things. A studio that takes a problem from formulation to store listing cannot take on many at once. In exchange, the things it does ship are coherent all the way down — the number on the screen means what the model says it means, and the model says something defensible.
Mathematics, artificial intelligence and data science are not the product. They are how the product gets to be right.
03Mission
Transform mathematics, artificial intelligence and data science into intelligent products that solve real-world problems.
Every word there is load-bearing. Transform, because a result is not yet a product. Real-world, because the constraint that makes a problem hard is usually the one that does not appear in the paper. And solve, because a product that merely displays data has left the actual work to the user.
04Vision
A portfolio of unrelated products, each built on a model we understand completely.
Not a company defined by one market. The through-line is the method: formulate, solve, ship, and keep the assumptions visible. A studio that can do that well in one domain can do it in the next, and that transferability is the only advantage worth compounding.
05Values
Six commitments we can actually be held to.
Values are only useful when they can be violated. These are written so that you could point at a specific decision and say we broke one.
Correctness before polish
A beautiful interface over a wrong number is worse than no product. The model is reviewed before the pixels are.
Explain the assumptions
Every model simplifies. We say which simplifications we made and where they break, inside the product itself — not only in the documentation.
The smallest thing that works
No framework, dependency or abstraction earns its place by being fashionable. It earns its place by removing more complexity than it adds.
Privacy is an architecture, not a policy
The strongest privacy guarantee is data that was never collected. We design for local-first and minimum collection before we write a privacy page about it.
Own the whole path
From the formulation to the store listing. Handing a model over the wall to someone else's product is how good mathematics ends up unused.
Say what is true
No invented benchmarks, no borrowed credibility, no roadmap presented as a shipped feature. Trust compounds and it is the only asset a young lab has.
06Method
Three stages, in this order, every time.
The sequence matters more than any individual technique. Skipping the first stage is the most expensive mistake available in this line of work.
Formulate
We start by writing the problem down precisely: variables, constraints, objective, and what would count as being wrong. Most failed projects fail here, quietly, months before anyone notices.
Model & solve
Then we choose the smallest machinery that can answer it — a closed-form model, an optimizer, a learned model, or occasionally a well-chosen heuristic. Complexity is a cost, never a credential.
Ship & measure
A result that never leaves a notebook has produced nothing. We build the product around the model, instrument it, and keep the assumptions visible to the people relying on them.
Work with us
The studio is small and selective.
That means we answer email personally, and that we say no to work we cannot do properly. If you have a problem that needs modelling rather than staffing, describe it and we will tell you honestly whether we are the right people.