For most of my career I was an unlikely candidate to build software. I am a consultant cardiac anesthesiologist. My world was the operating room, the bypass machine, the bleeding patient, the next case.

What changed wasn’t ambition. It was a quiet realization: Life had become loud and fast, and I no longer had time to sit down, open a textbook, and read the way I once did. And I started noticing the same pressure on the people around me, especially the younger anesthesiologists.

I watched them struggle with the questions that define our work. What is the right anesthetic plan for this patient, given this surgery, these comorbidities, these medications, these labs? Which technique fits this particular person? And underneath it all, the quiet dread every one of us knows, the fear of a catastrophe unfolding mid-case, and the uncertainty of how to respond in the seconds that matter.

I couldn’t give them more hours in the day. But maybe I could build them a colleague who was always available.

Building a companion, one gap at a time

I started with what frightened residents most: the anesthetic plan, and the crises. I built a module that takes the patient in front of you and offers a structured plan, and, when something goes wrong, a calm prompt toward management, with drugs and doses appropriate to that patient. Not to tell anyone what to do, but to be immediate support beside a clinician who, in that moment, has no one else to ask.

Then the next gap revealed itself, and the next. I added analysis of ABG results, ROTEM and TEG tracings, and PCA setup for postoperative pain. I asked why the tool shouldn’t help a clinician anticipate trouble rather than just react to it, so I added scoring systems to flag who might deteriorate. I added a logbook to record cases. I made it possible to build several anesthetic plans in advance and save them, ready to execute across a list, and to print them or send them to a colleague.

Finally, I came back to where it all started: the fact that I no longer had time to read. So I added a feature that lets a clinician type any topic and retrieves recent articles on it, deliberately drawing only from reputable, current medical sources the clinician can then read for themselves. Not to replace their reading, but to make it possible again inside a day that no longer leaves room for it.

The tracing in the resident’s hand

A parallel story was unfolding in the cardiac theater. One ordinary day a resident stopped me, a ROTEM tracing in his hand and an uncertain look on his face. We talked it through, he thanked me, he left. But the question stayed with me: why isn’t there a tool that reads this result and walks you toward the interpretation, and the next step?

Viscoelastic testing is exactly the moment where the data is in front of you, the patient is bleeding, and the distance between “I have the result” and “I know what to do” is where things go wrong, not because the answer is unknowable, but because no one is standing next to you at 2 a.m. to think it through.

So that became a tool of its own, focused on the cardiac and coagulation side. It read the ROTEM, then suggested management and weight-based dosing. I added camera input to ease the friction of typing values mid-case. I added TEG for hospitals without a ROTEM machine, then ABG and metabolic analysis, and from there, ventilation suggestions and an indirect-Fick estimate of cardiac output, using values the clinician already had. The last piece was ECMO, because I’d watched how often it stayed a black box for capable clinicians who simply approached it with hesitation.

Two tools, built years apart, from the same instinct: notice a colleague struggling, then refuse to walk around the problem.

The part I still can’t quite believe

I wrote up the cardiac tool and the evidence behind it and submitted it, a solo manuscript, a working anesthesiologist with no team behind him, to the Journal of Cardiothoracic and Vascular Anesthesia, the oldest subspecialty journal in anesthesiology, nearly forty years old. It was accepted.

Then an email arrived from Dr. Joel Kaplan, the journal’s founding editor-in-chief, a name on the spine of the textbooks I trained from. I read it more than once. I think of it as a medal on my chest. Not because it validated the code, but because it told me a working clinician’s idea, built at the bedside, was worth the field’s attention.

What I’d tell the next clinician with an itch

These tools now reach colleagues across roughly 150 countries. I’m proud of that, but the number isn’t the lesson. The lesson is that the best clinical software doesn’t come from technologists guessing at what we need. It comes from us, the people who feel the pressure of a fast day, who watch a junior colleague hesitate, who stand at the table when the tracing is ambiguous and the patient is bleeding.

A word I want to be exact about: These are decision-support tools. They elevate a clinician’s thinking, never replace the judgment only a human at the bedside can exercise. The dosing prompts are prompts. The literature it surfaces points to sources you read yourself. I designed for that line deliberately, because the goal was never to remove the colleague the resident came looking for. It was to make sure that when no colleague is within reach, and there’s no longer time to open the book, the silence has something useful in it.

If you’re a clinician with a problem that won’t let you go: Start with one piece. Build the smallest thing that solves the smallest version of it. Then listen to the next gap. That’s the whole method. It’s how a fast day and a resident’s question became something that traveled further than I ever imagined.

Ahmed Elsonbaty is a consultant cardiac anesthesiologist at King Fahd Armed Forces Hospital in Jeddah, Saudi Arabia, and assistant professor of anesthesiology at the Faculty of Medicine, Cairo University, Egypt.

Alongside his clinical practice, he builds free, clinician-designed decision-support tools used by anesthesiologists in roughly 150 countries. These include Anaesth-AI, a general anesthesia reference, and CCA-Pro, a cardiac and coagulation tool recently published in the Journal of Cardiothoracic and Vascular Anesthesia. His work focuses on closing the gap between clinical data and the next decision at the bedside, while keeping human judgment firmly at the center.

He writes about the experience of building clinical software as a practicing physician and shares updates on LinkedIn.