Two months, twenty years

When I was eleven, a doctor at Seoul National University Hospital gave my father two months to live. It was a late-stage brain tumor. My father was forty-two, and my mother was thirty-three.

He lived twenty more years.

After the surgery, visits to the intensive care unit were rationed to a few minutes, one visitor at a time. No one knew which visit might be the last. My mother gave one of her visits to me, so that her son could see his father at least once. She prepared me by saying, “There are wires in his head from the operation. Don't be startled. Take this towel and wipe his face, gently.”

I still remember the silence of that room, and my father's face under my hands.

The twenty years that followed did not move in a straight line. There was radiation, then recovery. Chemotherapy, then decline, then recovery again. A visiting nurse came to our home on a schedule; between her visits, while my mother was out earning a living for us, I was a young boy taking care of my very ill father.

My father kept the films from his scans as a keepsake of the surgery that had gone well; they lived in our closet for years. But the images showed him on a single day — the day of the operation. The twenty years that followed weren't anywhere in those scans.

Looking back as someone who now builds medical AI, I can finally name the nightmare my family was living inside of.

The two-month prognosis was a snapshot — the best single-point prediction available at the time — and it was off by a factor of a hundred and twenty. What actually determined my father's life was everything that happened after the scans: the surgery, the surveillance, the interventions, and the spaces in between them.

This was the first lesson I learned:

My father was never a prognosis. He was a trajectory.

In his early sixties, the trajectory turned for the last time, faster than anyone's observation could keep up with. His speech went, then parts of his cognition, then one side of his body. This led to long hospitalizations, and a nursing facility.

The deterioration didn't happen at a single checkup, or fit neatly inside a reporting cycle. The deterioration unfolded in the weeks in between these touchpoints.

I have devoted my career to responding to my family's story, one company at a time.

I joined Lunit (Mission: Conquering Cancer), because my father's cancer was found late, and I wanted to develop machines to find it early. I helped carry that mission through the company's public listing, and it continues giving back years of health to patients around the world.

Then, as CEO of Promedius, I went deeper with Healthy Aging through AI.We built the world's first chest-X-ray-based osteoporosis screening, finding a disease no one was looking for, in an image taken for another reason.

A cost-effectiveness analysis across six countries — Korea and the United States among them — projected roughly 1.8 million fractures prevented and some $50 billion in healthcare costs saved.*

It's easy to read a number like that as simply good healthcare economics. I read it as 1.8 million families. That was the second lesson: the most expensive patient in medicine is the one hiding in plain sight.

Yeji X is my third company, and it's the one the first two inspired me to build. Detection is no longer the bottleneck in medicine: timing and action are.

At Yeji X, we read the artifacts medicine produces: the labs, the medications, the imaging, day by day.

We model each patient the way my father was never modeled. We don't assign a score once, we put patients on a trajectory that can bend and flex. We start where the signal is richest, inside the hospital stay, flagging readmission risk before discharge instead of after the claim. And we model patients who look healthy on paper (as my father did between visits) so that intervention arrives before decline instead of after a bill arrives.

When this is rolled out across a nation of aging patients, two ledgers move at once: the clinical outcomes of families like mine, and the billions their health systems spend each year on outcomes that weren't prevented.

We named our company Yeji — 예지. In Korean, yejiis made of two words that share one sound: 豫知 — foresight, knowing what happens next — and 叡智 — wisdom, a clear and penetrating mind that sees into the nature of things. Yeji is the eye that sees what's coming, and the mind that knows what to do about it.

I named our company after the powerful force my family never had.

My father lived for an additional twenty years that no one predicted. Those years happened because of a brilliant surgeon, a devoted visiting nurse, a tireless mother, an eleven-year-old with a warm towel — and luck.

Our mission at Yeji X is Prevention through Intelligence, because the families that come after mine shouldn't rely on luck. They deserve to get those years by design.

David Jung · Founder & CEO, Yeji X · Nashville, TN · david.jung@yejix.ai

* Cost-effectiveness analysis across six countries (Korea, the U.S., Japan, Germany, Singapore, and Australia): approximately 1.8 million fractures prevented and ~$50 billion in projected healthcare savings at adoption. Presented at WCO-IOF-ESCEO 2026.