Mr Ahmd MHMD

About

Dr Amara Osei is one of Australia's most experienced cardiothoracic surgeons. Over a career spanning 24 years, he has performed more than 1,400 complex procedures — coronary artery bypass grafts, valve replacements, lung resections, and thoracic reconstructions — at major teaching hospitals across Australia. He currently serves as Head of Cardiothoracic Surgery at a leading Australian teaching hospital, where he leads a team of surgeons, perfusionists, and anaesthetic specialists.

Dr Osei completed his undergraduate medical training at the University of Melbourne, graduating with honours before undertaking specialist cardiothoracic surgical training through the Royal Australasian College of Surgeons. He subsequently pursued a PhD in surgical decision-making — an academic focus that would define the trajectory of his career. His doctoral research examined how experienced surgeons integrate incomplete and conflicting intraoperative data to reach consequential decisions under time pressure. The question that animated that work — why do two surgeons with the same training sometimes reach different decisions at the same critical juncture? — has continued to shape his research agenda.

Over the following decade, Dr Osei led research programs examining surgical risk stratification, hemodynamic instability patterns, and the cognitive architecture of expert surgical judgment. His work has been presented at international congresses and published in peer-reviewed journals covering both clinical surgery and surgical education. He holds adjunct academic appointments at two Australian medical schools, where he teaches senior residents and surgical registrars on the structured analysis of operative risk.

A recurring theme throughout his clinical and academic career has been the gap between what senior surgeons know and what they can articulate. The knowledge that separates a good surgeon from a great one is rarely codified. It lives in pattern recognition accumulated across thousands of cases — the slightly unusual rhythm of a de-airing step, the way a myocardium looks before it begins to fail, the particular haemodynamic trajectory that precedes a difficult separation from bypass. Junior surgeons acquire this knowledge slowly, through supervised practice and the accumulation of their own cases. Dr Osei joined Praxa because he believes this process can be accelerated — that tacit surgical expertise can be structured, encoded, and made available to the next generation of practitioners at a scale that mentorship alone cannot reach.

He brings to Praxa not only 24 years of operative experience but also a decade of structured academic work on how surgical expertise functions. That combination — practitioner knowledge and a rigorous framework for understanding it — is what makes his collaboration distinctive.

Domain Expertise

• Intraoperative decision-making under uncertainty — managing incomplete, conflicting, or deteriorating operative information
• Hemodynamic management — recognising and responding to haemodynamic instability during complex cardiac procedures
• Post-operative complication prediction — identifying patients at elevated risk before complications become clinically apparent
• Surgical risk stratification — pre-operative assessment frameworks for high-acuity cardiothoracic cases
• Team communication protocols in high-stress procedures — the structured language and anticipatory cues that define expert operative teams
• Separation from cardiopulmonary bypass — judgement regarding timing, support requirements, and failure patterns
• Cardioprotection strategy — myocardial protection decision-making across varying patient profiles and operative contexts

Why AI — and Why Now

Dr Osei has spent years watching medical AI tools fail in operating theatres. Not for lack of data. Not for lack of computational power. They fail because they are built on the assumption that clinical expertise can be derived from retrospective data alone — that if you analyse enough patient records, the patterns that distinguish expert decisions from novice ones will surface. They do not.

What distinguishes an expert surgeon is not primarily what they know about populations. It is what they perceive in a specific patient, in a specific operative context, at a specific moment — and what they choose to do about it. That perception is built from thousands of cases and shaped by years of feedback from experienced colleagues. It is tacit in the truest sense: difficult to articulate, impossible to extract from a medical record, and invisible to any algorithm trained only on outcomes.

The clinical decision support tools Dr Osei has encountered during his career have been designed, in the main, by people who have never stood at an operating table. They optimise for the measurable and miss the practitioner's judgment about what the measurable means in context. A pre-operative risk score can tell you that a patient is high-risk. It cannot tell you whether the anatomy of this specific case, in the hands of this specific team, on this specific day, warrants proceeding. That judgment belongs to the surgeon — and it is not currently captured anywhere.

What Praxa is building is different. The starting point is not the data. It is the practitioner. The product encodes the judgment of a surgeon who has developed, over more than two decades, an understanding of what the data means — and where the data is insufficient. That is the ingredient medical AI has been missing.

Collaboration with Praxa

Dr Osei's role in the Praxa product-building process is substantive and ongoing. He is not brought in to review a completed product and confirm it seems clinically reasonable. He is involved from the point at which the product is a set of questions: what does expert judgment in this domain actually look like? What decisions matter? What information do those decisions depend on? What does an expert surgeon attend to that a less experienced one misses?

Working with Praxa's product and engineering teams, Dr Osei structures his clinical knowledge — the decision frameworks, the pattern recognition heuristics, the contextual rules that govern when a standard protocol applies and when it does not — into a form that can be embedded in a working product. This process requires him to make explicit what has always been implicit, and to do so with a precision that holds up to technical scrutiny.

The result is a product that reflects not only what a cardiothoracic surgeon knows, but how a cardiothoracic surgeon thinks. Read more about Praxa's approach to domain-expert collaboration in our blog: Why domain expertise is the missing ingredient in AI product development.