Most diligence runs through a single lens. Every engagement here runs through three — clinical, commercial, and technical — for investors evaluating an asset and founders building one. Cardiovascular devices across the field; healthcare AI in any specialty.
One problem, examined three ways.
Will this hold up in a real patient?
MD, FACC, FSCAI — practicing cardiologist: interventional and structural heart
Does the model actually close?
MBA — Columbia Business School
Does it do what they say it does?
MS Biomedical Engineering — Johns Hopkins
Does the underlying claim hold, and will clinicians actually use it? Every engagement comes back to those two questions.
Devices across cardiovascular innovation — from anatomic thesis and mechanism of action through trial design, procedural reality, and clinician adoption. Read by someone who still does the cases, with interventional and structural heart as the clinical base.
Clinical, diagnostic, and workflow AI — not limited to cardiology. Assessed at the implementation level as well as the clinical one, so the review is not a physician taking the engineering team's word for it.
StructuraMed is the independent advisory practice of a practicing interventional and structural cardiologist with training in medicine (MD, FACC, FSCAI), business (MBA), and biomedical engineering (MS).
The clinical base is interventional and structural heart — including service as site principal investigator on more than ten pivotal device trials across TAVR, transcatheter mitral and tricuspid repair, and related therapies. That is where the procedural instinct comes from: what a device does in a real patient, on a real list, in hands that are not the inventor's.
The advisory work is deliberately wider than that base. Engagements run across cardiovascular innovation generally — transcatheter therapies, coronary and vascular devices, electrophysiology and lead management — because the questions that decide a product are the same everywhere: does the anatomy cooperate, does the physiology hold, and does the workflow survive a busy day.
On the AI side the work is hands-on rather than supervisory. Alongside the biomedical engineering background, that includes formal training in agentic AI systems through the Harvard Data Science Initiative and ongoing graduate coursework at Johns Hopkins at the programming level — so an AI review engages with how a system actually behaves, not only with how it is described.
Engagements serve both sides of the table: investors and venture teams evaluating an asset, and founders building one. Work is provided in an independent personal capacity.
Focused on the questions that matter before a product reaches its next stage.
Independent technical review of devices, AI tools, and clinical theses for investors, venture teams, and acquirers. Anatomic suitability, mechanism of action, model behaviour, trial design and endpoints, competitive and adoption realities.
Clinical advisory across cardiovascular innovation — structural, coronary, vascular access, electrophysiology and lead management, and adjacent areas. Anatomic, procedural, and adoption input from an operator who still does the cases.
Guidance for clinical, diagnostic, and workflow AI companies in any specialty. Clinical validity of the thesis, evidence and study strategy, workflow fit, and the real reasons clinicians adopt — or quietly abandon — new tools.
Founding clinical advisor and scientific or clinical advisory board seats for early- and growth-stage companies across both practices, with defined scope and disclosed compensation.
First-principles input on device design, procedural workflow, patient selection, endpoint construction, and overall clinical study strategy — for devices and for AI validation alike.
Faculty, KOL, and educational engagements on cardiovascular therapies, procedural technique, and the clinical application of AI in medicine.
Products rarely fail because the science was wrong. They fail because of something nobody asked early enough — while it was still cheap to answer. Every engagement works through the same four, whether the asset is a catheter or a model.
For a device: the anatomy it needs, the anatomy it will meet, and whether the physiologic claim survives the difference. For an AI tool: what the model is really keying on, and whether that holds outside the data it was trained on. Both questions are the same question.
Inclusion criteria, endpoint construction, and how enrolment actually behaved decide whether a result means anything outside the study. Having run sites on more than ten pivotal trials, the read is on what the protocol did — not only on what the paper reports.
The real-world adoption lens: what clinicians will actually do, and what they'll quietly resist. An extra step, an extra person in the room, extra radiation, or an unfamiliar motion — any one of them can end an otherwise excellent product, and none of them appear in the data.
Every engagement includes the case against. The most useful thing an independent reviewer brings is not enthusiasm — it is the objection the room has not raised yet, delivered while there is still time to act on it.
Scope and structure are agreed in writing before any work begins. Conflicts are disclosed up front.
A defined technical review of a single device asset or clinical thesis, delivered as a written memo with a live readout for the investment team.
For healthcare AI: clinical validity of the thesis, validation design, distribution shift and failure modes, regulatory evidence strategy, and workflow fit — engaging with the implementation, not only the deck.
Recurring clinical counsel for a company or fund — a standing call cadence plus ad-hoc review as questions arise.
Clinical or scientific advisory board seats, with defined time commitment and disclosed compensation.
Structure. Cash, equity, or a combination. Equity is considered for early-stage companies where the advisory relationship is substantive and ongoing — structured and disclosed like any other holding.
Notes on cardiovascular medicine, quantitative reasoning, and the clinical application of AI.
The fastest route is to book a call directly. For anything that needs detail first, send a note and I will respond personally.
A short introductory conversation to scope the question and confirm there is no conflict.
Choose a time