Co-founder · CTO
Architect of the IDS platform
18 years of software architecture. An industrial ERP he designed in 2012 is still running today — and he is the one rebuilding it.
The method's pilot
Our method rests on three pillars. Here they are — and here is what sits behind each one.
A 19-module SDK, started in 2022. Nine domain platforms built as a stable core with opt-in extensions: a client's specifics go through extension points and configuration — never through a diverging copy of the core. That is why the second client costs less than the first.
A module only becomes "industrialised" after two distinct client projects with no divergence from the core, and a "product" after three projects, code-free configuration and six months of stable releases. The rule is written down, dated, and can be held against us.
Framing, trade-offs and review stay human. Execution is delegated to agents — and constrained by checks that break the build. That machinery is what allowed nine domain platforms and five client missions to run in parallel with a small team.
The guardrails
Four checks, run on every integration. None of them is optional.
The business domain may not depend on infrastructure, and cycles between packages are forbidden. This is not a convention a review might wave through: it is a test that fails.
Every event's schema is generated from the code, published and verified automatically. An incompatible change cannot reach production unseen.
Mutation testing deliberately injects defects into the code: if the test suite stays green, it was testing nothing. High coverage proves little; this proves a lot.
Every endpoint is exercised by automated tests against a real database, and code quality must clear a blocking threshold before any merge.
Two stories
Designed in 2012 for a construction-industry manufacturer — while he was at a previous employer — this integrated management system is still at the heart of its operations: fourteen years in production without interruption. The relationship, now sixteen years old, is carried by INNOV DS today — and the rebuild into microservices is led by the person who wrote it.
On a consumer platform run alongside client work, he led the transformation of a monolith into a microservices architecture, and ran Kubernetes in production for seven years. Scaling is not a theoretical topic here: it was lived.
The track record
The milestones before 2020 predate the founding of INNOV DS: they were carried out at successive employers, or as a partner in a venture.
2008
Before INNOV DSJEE development, then team management — institutional intranets and business applications.
2012
Before INNOV DSGo-live of the industrial ERP he designed end to end.
2014
Before INNOV DSDocument management and a workflow engine, in a banking context.
2017
Before INNOV DSPartner and technical lead of a consumer platform — monolith to microservices.
2020
At INNOV DSCo-founder of INNOV DS, where he defines the technical core and the architecture.
2022
At INNOV DSFirst lines of the Java SDK that becomes the technical core of INNOV DS.
Today
Describe your process: you will get a straight answer on what is feasible, at what cost and in what time.
Talk about your project