The only non-physician in a family of doctors spent her childhood summers digitizing her father's paper charts. Two decades and one Morgan Stanley career later, she is building the AI that reads all of them.
Most of what a doctor knows about you is not in a tidy box on a screen. It is buried in a paragraph. A note scrawled after an appointment, a discharge summary, a line about how you responded to a drug three years ago. Computers, for decades, have been very good at the boxes and nearly blind to the paragraphs.
Eirini Schlosser built a company around the paragraphs. Dyania Health, which she founded in 2019 and runs as CEO, makes Synapsis AI - a natural language understanding platform trained to read unstructured electronic medical records the way a clinician would. It pre-screens patients for clinical trial eligibility. It automates the chart abstraction that normally has a human being sitting at a desk filling out a form line by line. It turns free text into something a research team can actually act on.
The pitch is deceptively plain. "Same technology," she has said, "but it saves hospitals millions of dollars where normally they would have to have a human being sitting there completing a form and doing manual chart review." The unglamorous work is the point. Medicine, in her words, is "an industry that has been quite honestly decades behind in terms of technological adoption versus other industries." She is not trying to reinvent the doctor. She is trying to give the doctor back the hours.
There is a hard constraint baked into all of it: trust. In clinical settings, an answer without a reason is useless. Physicians, she insists, need to understand "why the AI came up with that answer." So Synapsis AI is built to show its work, to point back at the sentence in the record that justified its conclusion. Explainability is not a feature she bolted on. It is the price of admission to the room.
It's a mix of madness, hard work, and luck, but I believe what combines these three is the persistence to keep going.— Eirini Schlosser
The strange specific: about twenty years ago, a teenager spent her summers feeding her father's paper patient records into a scanner. Not a metaphor for anything. An actual chore in an actual physician's office. "I spent my summers digitizing my father's patient records," she has said, "and grew up to be the only non-physician in my family."
She called herself the black sheep - the one who did not go into medicine. But she was the one who saw the records as data before anyone else in the house did. That vantage point, medicine viewed from just outside the profession, runs through everything she has built since. She understands the paperwork because she once was the paperwork.
Before the startups, there was a trading floor. Schlosser worked at Morgan Stanley on more than $60 billion in technology, pharmaceutical, and consumer deals. It was a career most people would keep.
She left it. The thing that pulled her out was the same thing she had noticed as a teenager: the most valuable information kept getting made without access to the richer, messier data locked in free text. So she founded Chuz, an NLP-powered recommendation engine, raised over $1M in seed funding, and learned how to build. Then, in 2019, she started Dyania Health and pointed the same obsession straight at medicine.
Unstructured EMR text - notes, summaries, free-form history - goes in, messy and human.
➞NLP trained with physicians reads the clinical meaning, not just the keywords.
➞It pre-screens for trial eligibility and auto-fills the chart-review forms a human once did by hand.
➞Every answer points back to the sentence that justified it. Explainable, or it doesn't ship.
Clinical research burns enormous human hours on tasks a machine can read faster - screening records, abstracting charts, reconciling real-world data. The bars below sketch where Synapsis AI aims its leverage. Illustrative, not audited - the shape of the problem, drawn to scale.
For years, the smart money in health tech went to the structured stuff - the fields, the codes, the numbers that slot cleanly into a database. It was easier. It was legible to a machine on day one. And it left most of the clinical story on the floor, because the story lives in the notes.
Schlosser bet the other way. She spent a decade leading multidisciplinary teams pointed at exactly one thing: building technology that extracts meaning from unstructured free text. It is a harder problem. Language is ambiguous, doctors write in shorthand, the same condition gets described five ways across five records. But it is also where the value hides, and she had known that since she was a teenager scanning her father's files.
The wager only works if the people who use it believe it. That is why she assembled a team that pairs NLP applied scientists with practicing physicians - a group credited with more than 800 academic publications between them. The scientists teach the machine to read. The doctors make sure it reads the way medicine actually reads, and keep it honest when it drifts.
Ask her how a company like this survives and the answer is not a growth chart. "It's a mix of madness, hard work, and luck," she says, "but I believe what combines these three is the persistence to keep going." Coming from someone who left a Morgan Stanley career and started over twice, it reads less like a motivational line and more like a field report.
Chuz came first - an NLP recommendation engine that raised more than $1M in seed funding and grew to a small team. It did not become Dyania, but it taught her how to build one: how to hire, how to raise, how to turn a thesis about language into a product people pay for. When she launched Dyania Health in 2019, she was not a finance analyst dabbling in tech. She was a second-time founder who had done the unglamorous reps.
The recognition has started to catch up with the work. Endeavor Greece picked Dyania for its Scale Up program in 2022. In 2025, Inc. named her to its Female Founders 500. The through-line from a summer scanning job to a national founders list is not tidy, but it is consistent: she keeps finding the value everyone else skips because it looks like too much work.
Spends summers digitizing her father's paper patient records.
Earns a BSBA in Finance from Ohio State's Fisher College of Business.
Works $60B+ in tech, pharma and consumer deals at Morgan Stanley.
Founds Chuz, an NLP recommendation engine; raises $1M+ seed.
Launches Dyania Health to bring language understanding to clinical research.
Dyania selected for Endeavor Greece's Scale Up program.
Closes Series A; total funding reaches $17.55M.
Named to Inc.'s Female Founders 500.
"I spent my summers digitizing my father's patient records and grew up to be the only non-physician in my family."
"An industry that has been quite honestly decades behind in terms of technological adoption versus other industries."
"It saves hospitals millions of dollars where normally they would have a human being sitting there completing a form and doing manual chart review."
"Physicians need to understand why the AI came up with that answer."
Schlosser on the Outliers podcast - on persistence, the outsider's view of medicine, and why the messy data won.
She is the only non-physician in a family full of doctors - and the one who saw the charts as data.
Her healthcare career literally began with a scanner and her dad's filing cabinet.
She walked away from a finance career built on $60B in deals to build from zero.
Synapsis AI hides a physics joke: type "OH," get back "IO."
Dyania keeps two homes - Jersey City and Athens - and one very publishing-heavy team.
Chuz came first. The $1M seed round taught her how to build before medicine did.