01
Medicine from First Principles
We start with the medical task, not with existing device categories. For selected clinical questions, for example, we investigate whether AI-assisted ultrasound can replace conventional X-ray imaging.
The first research project
One tablet. Many medical capabilities.
A multimodal diagnostic platform built from a standard Android tablet, an ultrasound probe and AI-assisted software. Our first proof of modular architecture.
What we are building
A platform, not another ultrasound machine.
The tablet uses its camera, microphone, audio and touchscreen. External sensors extend its medical capabilities; ultrasound is the first addition. ECG, EEG and other modalities may follow.
Hardware interfaces, data processing, interpretation, user experience and human expertise are developed as independent modules.
Components, not monoliths
Sensing
Interpretation
User experience
Data
Shared semantics
Human expertise
A complete, functioning system ↗
Why we are doing this
Two theses. One demonstrator.
02
Proving Composability
We demonstrate that independently developed modules can form a fully functional system — and then be reused, replaced and extended.
The concrete end-to-end proof
Diagnosis is our first use case.
Diagnostic assessment is easier to define than treatment or therapy, yet already provides practical value. Treatment may be added later.
Two diagnostic questions
AFractures — selected clinical questions using AI-assisted ultrasound
BLung and pleural examinations — a different diagnostic context
Both use cases are research proposals to be evaluated in appropriate clinical settings, not already validated substitutes for established procedures.
Multiple user roles
01Physician-oriented interface
02Nursing and emergency-service workflows
03Further roles depending on the specific setting
User guidance and professional authorization are distinct. The modules share the same diagnostic capabilities.
Development strategy
First, make it work. Then make it replaceable.
01
End-to-end proof
Assemble a complete diagnostic-assessment application from semantically described, independently developed modules.
02
Reuse
Implement a second diagnostic question and several user interfaces using shared components.
03
Substitutability
Replace at least one component with an independently developed implementation — without rewriting its consumers.
04
Open continuation
Publish modules, models, interfaces and tests so other teams can develop additional applications.
The initial consortium
We are looking for people
who can build it.
Research teams in diagnostics, ultrasound and sensing, medical AI, ontologies, agent-assisted software development and user interaction. Independent partners collaborate on one functional end-to-end proof.
Join the research consortium