She walked out of an MD/PhD program to make a bet most scientists wouldn't: that machines, fed enough human brain tissue, could find medicines faster than mice ever have.
The scientist who decided the lab and the algorithm belonged under one roof.
Alice Zhang runs Verge Genomics from South San Francisco, the stretch of bayfront where biotech reinvents itself every few years. Her company does something quietly radical for the field: it skips the mouse. Instead of testing ideas in animal and cell models that often fail to predict what happens in people, Verge starts with human tissue - thousands of samples of actual brain - and lets machine learning read the patterns inside.
The premise is simple to say and hard to do. Most diseases of the brain are not caused by a single broken gene. They are caused by hundreds acting together. Zhang's company maps that whole network from DNA, RNA and protein data, then hunts for drugs that hit the network rather than one node. That is how Verge surfaced PIKFYVE, a novel target for ALS, and pushed a candidate from research bench to human trial in about four years - a pace that makes seasoned drug developers raise an eyebrow.
In late 2025 the company hit the wall every drug developer dreads. Its lead ALS candidate, VRG50635, did not help patients in an early trial. Zhang did not pretend otherwise. She rebranded the company as Verge Labs, cut roughly nine in ten jobs, and pointed what remained toward a new business: selling the data and the insights - the drug targets, the patient populations, the biomarker predictions - to pharma partners who want them. It is a humbler shape for a company that once aimed to make its own blockbuster. It is also, in her telling, the part of science nobody puts on a pitch deck.
"A more reasonable expectation," she said of the setback, "is that the technology is going to go through some setbacks, but there's going to be really important learnings that can be fed back into the platform." Translation: the failure is data too.
The biggest journey for me as a founder was turning my inexperience from an insecurity to a superpower.- Alice Zhang
Zhang was born in Maryland to parents who had left China. Her father was an activist in the 1978 Democracy Wall Movement and was granted political asylum in the United States; her mother holds a master's in electrical engineering. The family's history with the country did not stay in the past. At twelve, on her first trip to China, Zhang was placed under house arrest and then expelled along with her parents.
She went back anyway. As a high school senior she traveled to Henan Province to work with children orphaned by a government-linked blood-selling scandal that had spread HIV through rural villages. The experience set a sentence she still uses to describe her work: to impact as many people as possible, with the greatest delta possible per person. That arithmetic - reach times depth - is why she eventually chose drugs over a single clinic.
At Princeton she studied molecular biology, graduated with honors, founded a campus chapter of Physicians for Human Rights, and ran experiments in a genomics lab that produced peer-reviewed papers. Then came UCLA's MD/PhD program and a Soros Fellowship in 2012. She was on the credentialed track to becoming exactly the kind of physician-scientist the program is built to produce.
And then she left.
Awarded the Paul & Daisy Soros Fellowship for New Americans to pursue her MD at UCLA, where she launched the university's first clinic offering medical evaluations for asylum seekers.
She leaves her MD/PhD program - shocking peers and professors - to co-found Verge, pairing machine learning with human-tissue biology under one roof.
Named to the class, as the bet on AI-driven drug discovery starts to draw attention and capital.
Named to Fortune's 40 Under 40 and Fierce Biotech's Fiercest Women in Life Sciences. Verge advances VRG50635, its AI-discovered ALS candidate targeting PIKFYVE, into clinical trials.
The ALS trial fails to show benefit. Zhang rebrands the company as Verge Labs, lays off about 90% of staff, and refocuses on monetizing its datasets and insights for pharma and biotech partners.
The hard part of AI in biology isn't clever code - it's clean, plentiful, human data. So she built labs to generate proprietary training data that keeps improving the models.
Map the hundreds of genes co-regulated in a disease, then find drugs that hit the whole network at once - rather than chasing one target in isolation.
Put computer scientists and biologists on one integrated team from day one. Verge's pitch was a platform and a wet lab in the same building.
People told her scientists and first-time founders can't be CEOs. She decided youth meant fewer assumptions to unlearn - and ran with it.
It's not an old drug and a new target, and not an old target and a new drug. It's a new target and new drug.- On Verge's PIKFYVE discovery
Neurodegenerative disease has been a graveyard for drug programs. Zhang's argument is that the old tools - animal models, single-target screens - were never built for it. Verge's focus areas read like the hardest problems in medicine.
In Verge's scrappy early days she ran more than 1,000 interviews herself and called herself the company's "chief PowerPoint officer."
She keeps a weekly digital sabbath - a full day offline - and practices daily exercise and yoga.
She cooks and haunts farmer's markets, and admits she doesn't travel as much as she'd like.
Her aspiration by 40: deliver the first drug discovered by a computational platform into a human, and prove it works.
"A lot of people told me scientists and first-time founders can't be CEOs."
She built the company anyway.