About
A layer that didn't exist.
Axiome operates at one boundary: the point where a bioinformatics pipeline emits a result table and a scientist begins to interpret it. Below that line there are pipelines, data platforms and analysis environments, all mature. Above it there are institutional decisions. Between them, the step where a result becomes a conclusion has no infrastructure at all: it happens in scripts, slides, email and memory, and it does not survive.
Axiome is a French deeptech SME in Marseille, established March 2026. A first version is built and available for demonstration.
Where this goes next
A conclusion should survive its study.
A conclusion validated once should be reusable. Today, when a team faces new data, there is no reliable way to know whether a comparable conclusion was already reached, or whether it still holds. Conclusions drift and contradict each other without anyone noticing.
Axiome's research direction is to make a validated interpretation carry the context it depends on, so that applying it to new data becomes an explicit verdict rather than a judgement call nobody wants to sign.
Today, Axiome records and carries the conclusion forward. It does not tell you whether a past conclusion still applies to new data.
Research in progress, developed with academic partners. Not part of what a pilot buys today.
Why immuno-oncology
Where Axiome is built and tested.
Immuno-oncology is where Axiome is built, with the teams who work in it. It's the field where interpretation is hardest, where the data is largest, and where traceability has stopped being good practice and become an infrastructure requirement. If the mechanism holds here, it holds anywhere, but it is here that it has to hold first.
Marseille
A national immunology reference: CIML, CRCM, ICI, MIB, Cancéropôle PACA on the research side; AP-HM and Institut Paoli-Calmettes for care and cohorts; Marseille Immunopôle and Eurobiomed on the industry side. Validation happens in person, at high frequency.
Immuno-oncology first. Everything else follows from getting this right.
Founder
Who's building it.
Felipe Machado Guimarães
Founder
Twenty-two years in software, most of it in R&D, building things that didn't exist yet. Five of those at Hewlett-Packard, leading the interface of a high-performance remote graphics product across teams in the United States, Brazil and France. Since then I've led distributed engineering teams across Europe and the Americas.
Then diagnostics. At HalioDx I created the engineering team that took the Immunoscore image-analysis platform to CE-IVD marking and into clinical use in Europe, the Americas and China (ISO 13485, IEC 62304, and the audits that come with them), and became Associate Director of Software Engineering there. After the Veracyte acquisition I delivered a cloud-native multi-omics analytics platform for biomarker discovery, companion diagnostics and clinical trials.
Axiome comes from what I watched research teams actually spend their time on. Every chart, every assumption, every conclusion becomes a node in a connected body of work rather than a file in a folder, so the team can work in one place and the work is still there when someone asks about it later.
Research use. Axiome produces the interpretive record and the evidence pack. It never carries a device class, never outputs a clinical decision, and does not certify anything.
I defer to the experts on the biology, the methodology and the statistics. What I can speak to is what happens to a conclusion after someone draws it.
Advisors are named in pilot conversations.
Fit
Who this is for.
- Translational research laboratories, academic clusters and core facilities.
- Biotech and pharmaceutical teams working in immuno-oncology or multi-omics.
- Commercial cooperation across the EU, including paid pilots and reference sites.
The same workspace, seen from each of these seats, is set out in the use cases.
Start with your own data.
We can work from your existing pipeline outputs, in your infrastructure. We can sign an NDA before anything is shared.