ABOUT OMNIMEM

A doctor, a mathematician, a developer and a team, working on the same four problems

Omnimem is a Danish company in Ringe, on Funen. We build knowledge bases out of the material an organisation already holds, and the things you can do once you have one.

We are small on purpose. Every engagement is run by the people who built the system, which limits how many we can take at once and is the reason the answer to a difficult question does not have to travel through three layers to reach somebody who knows.

The work sits where clinical judgement, mathematics and working software meet, which is an awkward place for a company to sit and the reason this one exists. Health evidence is not a search problem or a modelling problem or a software problem. It is all three at once, and the failures happen where they meet.

The three of us cover those three things between us, and we argue about the overlap. That is where the design comes from.

WHO WE ARE

The people you will actually be working with

Not an account manager, and not a team you meet once at the pitch. These three are on every engagement, which is both the point and the limit.

Torben R. Uhrenholt

Chief executive · MD, PhD

A physician who kept meeting the same problem from the clinical side. The evidence needed to answer a question existed. Getting to it in the time available did not.

Chief Scientific Officer of Movement Health Foundation, where the health economics modelling in Ghana and Peru sits, and co-owner of SundVision, where the cancer plain-language work was built. He is the one who decides where a machine stops and a person starts, and he is difficult about it.

Andreas C. Uhrenholt

Co-founder · Applied mathematics

Applied mathematics, University of Southern Denmark. He is responsible for the part most people skip: how the things inside a body of material are made to relate to one another, and what happens to uncertainty as it travels through a model on its way to a number somebody will act on.

If a comparison table across fourteen countries holds together, or a health economic model behaves sensibly when a parameter is uncertain, that is his work.

Simon C. Uhrenholt

Co-founder · AI systems and development

Builds the thing the other two argue about. Retrieval, provenance, and the unglamorous work of making all of it run inside somebody else’s constraints: a ministry network, a hospital’s rules, or a sealed case with no connection at all.

Most of what a customer experiences as reliability is decisions he made about what the system does when something is missing.

THE OVERLAP

Why it takes all three

The four things we build look unrelated from outside. Underneath they are the same capability with the weight moved. What moves with it is which discipline has to be right.

A retrieval system a clinician would not sign off on is a demo. A clinically sound answer nobody can trace is an opinion. A model with excellent mathematics resting on parameters nobody accepted is a number waiting to be challenged. The failures are always at the joins, and the joins are why we work this way.

THE COMPANY

How we work, and what we turn down

We take a small number of engagements and stay on them. There is no separate delivery team, no handover, and nobody whose job is to manage the relationship rather than the work.

That has a cost, and it is worth being honest about it. We cannot take everything, we are not the fastest route to a proof of concept, and if what you need is a large team on site next month we are the wrong company.

What we turn down

Work where nobody will own the answer. If no person is willing to accept what the system produces, the system is decoration and the project will not survive its first disagreement.
Corpora nobody is willing to define. If what is in scope cannot be written down, the answers cannot be defended, and we would rather say that at the start.
Clinical decisions we think should stay with a clinician. There are questions we decline to automate. Which ones is a conversation we have before an agreement, not after.
Projects that need the evidence to say something in particular. We build systems that report what the material says. If that is inconvenient, this is the wrong tool.

Who checks our work

We are a small company making claims about accuracy in a field where being wrong matters. That is only worth anything if somebody outside the company tests it.

University of Southern Denmark
The Research Unit for General Practice has independently evaluated our clinical work, including the plain-language cancer material with patients and the offline clinical device. Evaluation designed and run by them, not by us.
Copenhagen Institute for Futures Studies
Our partner on knowledge base work for international health policy programmes, and the authors of the foresight work behind the offline device.
Movement Health Foundation
The health economics modelling programme our evidence architecture sits underneath, applied in Ghana and Peru.
Innovationsfonden
Public InnoBooster funding has supported both the offline clinical device and the cancer plain-language work, each with the reporting that comes with it.

FOR THE RECORD

The company details

Omnimem ApS
Lombjergevej 1
DK-5750 Ringe, Denmark

CVR DK42361453
support@omnimem.io

Bring the difficult question

The useful first conversation is not a capability demonstration. It is the question your organisation keeps answering by hand, and whether the material that would answer it already exists somewhere in your files.

Why Omnimem