Jenny Shao Left a White Coat for a Memory Machine

Jenny Shao’s founder story begins with an appliance behaving badly. For a middle-school science fair, she needed to incubate bacteria; the family oven ran too hot. The detail survives in a Cleveland Clinic profile because it is charming, but it is also useful. Long before venture funding or artificial intelligence entered the plot, Shao was already doing the essential work of a builder: finding the constraint, refusing its authority, and improvising a way around it.

The path that followed looked unusually tidy from the outside. She went to Columbia University, where she studied biochemistry and neuroscience and behavior. She spent three years researching learning and memory in the laboratory of Nobel laureate Eric Kandel. She completed medical school at Cleveland Clinic Lerner College of Medicine, trained in surgery at Mount Sinai, and later joined a Harvard residency program. Each institution made the next line on the résumé easier to predict.

Then, in February 2024, Shao made the prediction useless. She left residency, booked a one-way ticket from Boston to Silicon Valley, and founded the company now known as Robyn AI. Her first inventory was austere: no team, no product, no roadmap. What she did have was a stubborn thesis assembled across years of research and training. Computers had become brilliant at retrieving information. They remained clumsy at carrying the meaning of a relationship from one conversation into the next.

“We want to be Robin to everyone’s Batman.”Jenny Shao on the name behind the company

The useful baggage

Career pivots are usually narrated as bonfires. The old identity burns; the new one rises dramatically from the ash, preferably wearing a hoodie. Shao’s shift is more interesting because very little was discarded. The research became a technical lens. The clinical training became a discipline of boundaries and responsibility. The years spent publishing and organizing became practice in recruiting people around an idea.

At Columbia, she did more than work at the bench. As executive editor of the university’s Journal of Global Health, she helped grow a small campus operation into a network spanning 40 institutions, 10 countries, and five continents. Institution-building arrived before company-building. So did an appetite for scale.

Her research record was already substantial by the start of her Harvard training: more than 65 publications and presentations, including 22 peer-reviewed papers and five book chapters, according to an institutional biography. One paper lists her among a large interdisciplinary team developing a system for profiling neuron activity. Another early project used MATLAB to model tumor-immune dynamics and earned her recognition as an Intel Science Talent Search semifinalist. The subjects varied. The habit did not: define a complicated system, find a measurable pattern, and test whether the model holds.

Memory is not a filing cabinet

Shao’s central observation is deceptively plain: memory is not a folder where facts wait in alphabetical order. Human memory is revised by relevance. A detail can grow more important after a later experience. Emotion changes what sticks. Shared history turns a familiar sentence into a private joke, a warning, or a kindness. A good friend does not merely know more facts about you than a stranger. The friend knows which facts matter now.

That distinction became Robyn’s product argument. A conventional assistant can hold the current conversation and retrieve stored details. Shao wants relational software to preserve continuity, notice evolving patterns, and carry context forward. She has described the company’s architecture as a relational memory layer, borrowing a conceptual map from the neuroscience she studied years earlier.

The important word is relational. A stopwatch remembers a lap time. A calendar remembers an appointment. Neither remembers in a way that changes the character of the next exchange. Shao’s bet is that an assistant becomes more useful when the accumulated past alters how it responds in the present.

There is also a neat inversion in the company name. Robyn is the sidekick. The person using it gets the cape. Shao has said she chose the name because she believes everybody has a superhero inside, and a well-designed assistant should help that person emerge. It is a bit of comic-book whimsy carrying a serious product rule: the software does not get to steal the scene.

The one-way ticket

Leaving medicine makes for an irresistible headline because the sacrifice is legible. The years of education are visible. The prestige is visible. The uncertain value of an unbuilt product is, by definition, invisible. Yet Shao’s own account makes the move sound less like a revolt than a transfer of obligation. After a 24-hour surgical call, she concluded that the kind of connection she wanted to make available would require a different vehicle.

She took nine months to make the decision. That interval matters. Founder mythology likes the jump and edits out the ledge. Shao examined the mismatch, tested her conviction, and only then moved west. The one-way plane ticket was cinematic. The months before it were the work.

Robyn began with Shao alone and reached three people by the start of 2025. By the public launch that November, the company had ten. Shao’s hiring language is revealing. She says she looks for partners rather than employees: people with enough ownership to fix what is broken without waiting for permission and enough conviction to bring strong opinions into the room. The description could double as a portrait of the founder herself.

“The answer is simple: the right people.”Jenny Shao on what powered Robyn’s early growth

The public launch arrived with $5.5 million in seed funding led by M13. The round included a curious table of builders and operators: Google Maps co-founder Lars Rasmussen, early Canva investor Bill Tai, former Yahoo chief financial officer Ken Goldman, xAI co-founder Christian Szegedy, Sangha Capital, and ACTAI Ventures. Different résumés, one shared wager that the next interface shift may depend less on a machine’s fluency than on its capacity for context.

The discipline of a sidekick

The market around personal AI is crowded with personalities. Many products perform companionship by being endlessly agreeable or theatrically alive. Shao’s framing is narrower and, in a way, more demanding. Robyn is supposed to support reflection and connection while remaining clear about the limits of software. She has repeatedly argued that AI should empower people rather than replace them.

That boundary is more than positioning. It determines what the product should refuse, how it should handle sensitive conversations, and whether its success is measured by time spent with the machine or by what the person carries back into life. A sidekick must be useful without becoming the protagonist. Product people may recognize this as the harder design brief.

Shao’s scientific background gives her a language for the machinery, but her public metaphors are resolutely human. She talks about being seen, known, and remembered. She talks about a superhero in everyone. She talks about the team debugging at 3 a.m. and laughing together at game nights. The vocabulary shifts easily from synapses to sidekicks, from memory consolidation to Batman. The playfulness is strategic. Abstract technology becomes legible when people can find themselves inside the metaphor.

By early 2026, Shao said Robyn had logged hundreds of thousands of conversations and reached users across much of the world. Those numbers are evidence of curiosity, not a verdict. The harder questions remain ahead: whether persistent context consistently produces genuine usefulness, whether a young company can maintain its boundaries while it grows, and whether people will trust one more piece of software with the story of who they are becoming.

The old map, folded differently

Shao’s career is often described as a sharp turn: physician becomes founder. Stand far enough back and it looks more like a continuous line. At thirteen, she was drawn toward science by the mysteries that living systems refuse to explain neatly. At Columbia, she moved from biochemistry toward the brain. In Kandel’s lab, memory became the subject. In publishing, she learned to extend a small institution beyond one campus. In medicine, she encountered the emotional weight people carry through consequential moments. In Silicon Valley, those strands were tied into a company.

The lesson available for stealing is refreshingly unromantic. Your prior life does not need to be escaped for a new one to begin. It can be mined. Domain expertise becomes more valuable when translated, and translation is not the same as repetition. Shao did not open a clinic with nicer software. She took a model of human memory and asked what it could teach a machine about continuity.

Plenty of technology can remember what you said. Shao is building toward the more interesting standard: remembering why it mattered. The distance between those two capabilities is where Robyn now lives. It is also where the former researcher, former resident, and current founder has placed her most public bet.