— Ventures
Building,
not just advising.
Consulting funds the runway: capped, fixed-scope, and chosen so it never competes with the venture for attention. Dei Corpus is where I carry the outcome myself.
I build ventures on the same foundations I build client systems on: deep tech with a defensible technical wedge, in markets where the hard part is engineering rather than distribution. I am not interested in businesses whose main challenge is buying attention. I am interested in problems where the reason nobody has solved them is that solving them is technically difficult.
What I look for before I build.
A technical wedge that compounds
The first version has to be hard to copy and the tenth version harder still. If the moat is a feature, there is no moat.
A market where engineering is the constraint
Some markets are won purely by distribution and capital. I build where a hard technical wedge is the price of entry, knowing that in regulated markets, evidence and adoption are the second and third walls, and the roadmap has to budget for all three.
Regulatory literacy as an advantage
Medical AI, fintech, and crypto share a property: regulation eliminates most competitors before the product does. Understanding MDR, MiCA, MiFID II, and GDPR at implementation level is a durable advantage, not overhead.
A path to evidence, not just a path to launch
In deep tech, the milestone that matters is validation, not release. Every venture I build has a defined evidence roadmap before it has a marketing plan.
Dei Corpus
Multi-modal biomarker fusion for early disease detection.
The problem
Cardiovascular, metabolic, and dermatological signals are captured and read in isolation today. A cardiologist reads the ECG. An endocrinologist reads the glucose curve. A dermatologist reads the imaging. No system correlates them into a single, continuously updated risk picture.
This matters because the earliest warning patterns for several major disease categories emerge across domains rather than within one. A metabolic drift that is unremarkable on its own becomes significant alongside a specific cardiac rhythm variation. Read separately, both readings are normal. Read together, they are not.
The approach
Dei Corpus ingests structured physiological signals from already-certified sensors and applies a fusion architecture that correlates them across domains, producing a composite risk score designed for clinical interpretability rather than black-box scoring.
Building on CE-marked sensors rather than developing hardware is a deliberate decision: it removes hardware and manufacturing risk from the critical path and puts the defensibility where it belongs, in the fusion layer. The software itself will require its own conformity route. I expect classification as an MDR Class IIa medical device and a high-risk system under the EU AI Act, and the roadmap is built around that evidence pathway rather than around avoiding it.
Where it stands
TRL 4. End-to-end feasibility demonstrated on synthetic datasets modelled on published biomarker distributions. Synthetic data proves the pipeline, not the signal, which is exactly why the next phase is benchmarking on real-world datasets, PhysioNet first, while securing a Portuguese clinical partner for a pilot pathway in parallel.
Roadmap
For investors and partners.
I am open to conversations with investors, operating partners, and technical co-founders on Dei Corpus. This page is informational and is not an offer or a solicitation of investment. Specifics are discussed directly, under NDA where appropriate.
The most useful first message tells me what you typically back at what stage, and what you would want to see before going further. That gets us past the first three emails in one.