From Statistics to Education to Clinical Trials: The Entrepreneurial Journey of Will Ma (HopeAI)
Conversation: Xiao He × HopeAI Founder Will Ma
Introduction: From Academia to Pharma
Xiao:
Could you start by introducing yourself to our readers?
Will:
Sure. My name is Will. I did my undergraduate studies in mathematics at Zhejiang University, and then pursued a PhD in statistics at the University of Virginia. My doctoral research was on clinical trial design—things like response-adaptive designs. So naturally, after graduating, I joined the pharmaceutical industry. At Sanofi in New Jersey, I worked on two large Phase III clinical trials.
The first trial had a huge impact on me. Just four months into my job, the trial was terminated. By then, 60–70 million dollars had already been invested. The problem wasn’t the drug itself—it was miscommunication and misunderstandings between statisticians and clinicians. This is something the outside world rarely sees, but inside pharma, it’s not unusual. Clinical trials are all about making quick decisions under uncertainty, racing against time. If you’re too slow, failure becomes even more likely. And my very first trial “failed” like that.
From Pharma to Academia, Then to Entrepreneurship
After two years in pharma, I felt discouraged and went back to academia, taking a faculty position at the Moffitt Cancer Center (partly because one of my PhD advisors had just become department chair there). The research fit perfectly with my background.
Within two years I had published enough papers, secured my green card, and then officially jumped into full-time entrepreneurship—this time in the field of AI in education, specifically adaptive learning. Most of our business and team were in mainland China, with a small team in New York. We grew into one of the leading companies in China, and actually globally as well.
Later, due to a series of regulatory shifts, that entire sector in China was shut down. We were fortunate to pull out in time instead of struggling on—otherwise survival would have been nearly impossible. So we returned to the U.S. and went back to our roots in healthcare.
After reacquainting myself with pharma for a while, we restarted entrepreneurship in 2023, this time focusing on AI in clinical development. Many of my former core team members—CTO, operations lead, etc.—rejoined. I also invited my PhD advisor and friends from pharma to join. That’s how HopeAI began.
Academia and Entrepreneurship: Fate or Choice?
Xiao:
I noticed you even held a faculty position. Looking back now, how do you see the contrast between academia and entrepreneurship?
Will:
Sometimes you just have to believe in fate. People told me from a young age I’d become a teacher, and I never believed it. But after two years in pharma, I ended up back as a professor. Later, when I started an education company, it felt even more like “teaching.” Honestly, I do long for school life. Someday, I might still return to teach or write textbooks—that would be a beautiful ending.
Xiao:
Entrepreneurs are usually imagined as being out pitching clients and chasing deals, which feels so different from academia. Yet you’ve succeeded in startups but still long for the academic world?
Will:
It’s really about mindset at different stages. Doing research for its own sake is one thing right after graduation; returning to academia after industry is another. I just like experiencing life. I’ve done two years in industry, two years in academia, then startups, then pivots… In pharma, I quickly saw the ceiling: colleagues 20–30 years older were doing the same work as me, and I realized in 30 years I’d likely still be in the same place. And very few people rise to VP. Many end up contractors. That’s when I decided to change environments. Academia, especially with a family and kids, was much more supportive.
Why Education? Market and Timing
Xiao:
After getting your green card, why did you choose to start an education company in China instead of doing healthcare entrepreneurship in the U.S.?
Will:
This was around 2016. “AI” wasn’t as hot as today; people talked more about the internet. We looked at which big sectors the internet had barely touched: healthcare and education. Back then, both were almost untouched. Healthcare was extremely hard—my adaptive design research wasn’t mature enough to be applied in medicine. Education, on the other hand, was easier: giving a student a slightly better or worse problem set only affected efficiency, not life or death.
Second, China had the largest after-school tutoring market in the world. In the U.S., families spend about 30% of disposable income on mortgages and similar expenses. In China, many families put 30% into tutoring. The market was massive, with abundant cash flow, yet it was still at a “slash-and-burn” stage: low digitization, uneven teacher quality, huge inefficiencies. Bringing advanced AI into that environment was like a dimensionality-reduction strike. It was the right time, the right market.
On Choosing Co-Founders
Xiao:
Many aspiring founders wonder: how do you find good co-founders?
Will:
Most of my co-founders are close friends. It’s hard to imagine partnering with strangers. You spend more time with your co-founders than with your spouse. You date for years before marriage—co-founding requires even deeper trust, built over many years of collaboration and aligned values. HopeAI’s founding team includes my PhD advisor, my CTO and COO whom I’ve worked with for 7–10 years. Of course, conflicts happen—co-founder breakups are the No. 1 startup killer.
Xiao:
I’m impressed that your team has stayed intact—even when shifting from education to clinical trials. How did you make that possible?
Will:
Maybe it’s my “teacher” side. I like sharing and growing together. Moving from education to drug development is a huge leap. I spent two years systematically teaching my CTO and COO the entire drug development process—weekly classes, constant discussions. We also brought in my PhD advisor and industry experts to co-teach. The principle was: train the old team, don’t just swap them out.
Startup Culture: Optimism, Persistence, and Openness
Xiao:
When we last spoke, it felt like a crash course in entrepreneurship. You’re very open with your insights, while many people tend to be guarded.
Will:
We’re generally very open, even with competitors. Being willing to request and to share is key. Entrepreneurship often comes down to one decision, one conversation, one new idea that makes the difference. I’ve gone through four or five moments when payroll was nearly impossible—then suddenly, a turnaround appeared.
Xiao:
You are always cheerful. Are you naturally optimistic?
Will:
Our culture rests on four pillars:
Freedom & Responsibility (borrowed from Netflix—we actually studied their culture directly).
Optimism.
Persistence.
Persistence is essential, but you can’t persist if you’re miserable all the time. Even if we might not make payroll tomorrow, I still sleep. You take it seriously, but you also treat it like “the game of life.” If the company fails, that too is a meaningful experience.
Family and Entrepreneurship
Xiao:
What did your wife and parents think when you left academia to start a company?
Will:
My wife was against it at first. We were living in a low-cost area, I had a tenure-track job with good pay and light pressure, the kids liked it too—it was stable. But I didn’t want to “settle.” She realized she couldn’t stop me and reluctantly supported me. My parents were fine with it; they were happy I could spend more time in China.
Work and family are reflections of each other. How someone is at home is usually how they are at work. For founders, family support is critical—almost like marriage. Both families have to accept the long hours, the travel, the sacrifices.
Recommendations and Vision: From The Matrix to Healthcare AGI
Xiao:
To wrap up, could you recommend a book, podcast, or movie to our readers?
Will:
I’ll start with a movie: The Matrix. I’ve loved it since high school—it probably planted the seed for my interest in AI. What we’re doing in clinical trials is essentially simulation—running a virtual trial before the real one, to see if a design works.
Our vision is to build AGI for healthcare. Right now we focus on clinical trials, particularly “covariate-adjusted response adaptive design.” Instead of a 1:1 randomization where some patients get lucky and some don’t, we aim to assign each patient to the treatment most suitable for them, while still preserving necessary randomness. We’ve researched this for years, and now it’s ready to be applied.
Looking ahead, we believe the world will converge into one giant “clinical trial”: the real world. The healthcare system itself becomes the algorithm. Instead of clinical trials searching for patients, everything starts from the patient: based on their profile, the system decides whether they should get an approved drug A, drug B, or an experimental drug C, and assigns probabilities accordingly. It’s like The Matrix, connecting all patients and all drugs in one system.
People often say, “Life-and-death decisions can’t be left to AI.” I believe the opposite: in the next decade, life-and-death decisions won’t be trusted to humans. Driving, treatment decisions—humans can’t shoulder that responsibility. Only AI or structured systems can.
So we’re starting with structuring and making clinical evidence usable. Once evidence is AI-ready, treatment and trial design can be based on it. That’s our mission.
As for books:
Google’s management playbook (I learned OKRs like a textbook when I first started).
Netflix’s culture handbook (we studied their culture).
Andy Grove’s High Output Management (a must-read for us).
Paul Graham’s essays.
During my years in China, I also studied Alibaba’s sales methodology. In the early 2000s, they convinced small businesses to pay thousands of dollars to list online—when few even understood the internet. That was an incredibly tough sales job, and they developed a deep system around it. We spent years learning from that.
Sales cultures differ sharply between China and the U.S.:
In China, half of business runs on value, half on “hype.” Salespeople may not be technical but must close deals.
In the U.S., sales is value-driven, product-driven. Pharma has a finite number of clients, and every deal matters. Founder-led sales work better here.
At one point, we even physically separated sales and R&D: two buildings, no interference. Free-form Silicon Valley culture doesn’t work for sales, and KPI-heavy sales culture destroys engineers. We had ~150–200 staff: 60–70 in sales and front-end, the rest in R&D. Separation kept the cultures healthy.
Xiao:
So HopeAI’s sales are still mostly founder-led?
Will:
Yes. In the U.S., expertise is crucial—there’s no such thing as a salesperson who can sell without understanding the business. There are only a handful of large pharma companies, and each opportunity is precious. Once you establish enough collaborations, smaller companies see your team’s credibility and value and want to work with you. Doing business in the U.S. is more “pure”.
Xiao:
Thank you so much, Will. This was incredibly insightful.
Will:
Thank you, Xiao!
