Execudo/

We design and build intelligent software. Agents, platforms, ventures. A small Paris studio, from idea to production.

Start building
01 / The thesis

Models are an ingredient.
The work is software /

AI matters when it changes the work. The craft is to turn a model into a system people can trust: product surface, tools, data, infra, feedback, all forged into one clean machine.

02 / The work

What we build/

An idea might become an app, a system, a tool, or a company operated by software.

We start from the problem, not the model. We find the right shape for it, then design and build it end to end: short cycles, real software in production, infrastructure we own.

Fig.1An app

A focused tool, shipped to production. One job, done precisely, with no dashboard nobody asked for.

React · API · shipped

Fig.2A system

Several workflows, composed into one. It watches, decides, and acts on its own, and asks you only when it should.

Workflows · agents · autonomous

Fig.3A company

A venture where the software runs the business, not the slides. Built AI-native from the first line.

Software-operated venture

03 / The method

How we think/

We keep the stack understandable: model choice, product surface, data flow, deployment, feedback. The goal is simple, software that keeps working when nobody is watching.

Vertical/beats horizontal

We choose a narrow problem and build the full path around it. The interface, the agent, the data, the alerts, the boring parts too.

Production/beats prototype

A demo is a sketch. We care about the thing that survives users, errors, latency, invoices, and Monday morning.

Software/beats headcount

If the same work repeats, the answer is often a system: agent, queue, script, interface, log. Small pieces, wired well.

04 / The studio

A small studio in Paris, building systems.

We are young enough to move fast, serious enough to care about what happens after the demo. Product, design, infra, open-source tools, model evals, deployment, we like the whole system.

Small by design/

Small team, short loops, direct ownership. No layers between the idea and the build.

AI-native/

The model is not decoration. It sits inside the workflow, with tools, memory, checks, and human judgement.

Owned stack/

We prefer systems we can understand, host, debug, and improve. Control is part of the product.

Long view/

AI will change how companies operate. We build small enough to ship, deep enough to matter.

05 / Field notes

Field notes/

A quiet log of what we are testing, reading, and learning while building intelligent software.

06 / Contact

Bring us
a problem /