The room was mostly PhD students, postdocs, facility managers, and scientists. People who may or may not have started a company but were curious about what the journey actually looks like from the inside. And what surprised me was how many of them stayed after to keep talking. The questions didn’t stop.
We ran it under Chatham House rules so the panelists could be real. They were. We recorded it because we’d run the same format at ASMS earlier in the year and the feedback was: “We couldn’t scribble fast enough.” This time we came prepared.
Here’s what came out of that room. And honestly, most of it applies whether you ever start a company or not.
From left to right below:
Lindsay Pino, Co-Founder and CTO of Talus Bioscience
Baljit Ubhi, CCO at Matterworks
Five very different paths. But one thing kept coming up: the skills you’re building as a scientist right now are more transferable than you think.
When I asked what launched each person into entrepreneurship, Parag’s answer wasn’t what anyone expected.
He was running his lab at Stanford. Things were fine but he was restless. He decided to do a classic American road trip to clear his head. Except the one-way rental from New York to California was too expensive. So he wrote a program to find the cheapest one-way rental in the country. Detroit to Denver. Two and a half weeks. No plan.
Every morning he picked a direction and drove.
The weekend he got home, he woke up and the first words out of his mouth were: “Oh, that’s how you do it.” His wife looked at him, puzzled. He spent the rest of the weekend talking to himself and running simulations.
By Monday he had the seed of what would become Nautilus: a way to do proteomics without a mass spectrometer. From a person whose lab had six of them.
Here’s the part that matters for anyone sitting in a lab right now wondering “could I do this?” Parag didn’t start with “I want to build a company.” He started with “I have an idea I think is important.” The company was the vehicle that could move fast enough to test it.
He looked at the options. Build it in his academic lab? Pitch it to one of the big instrument companies? Start a company? He chose the startup because it was the only vehicle that could iterate quickly enough on something that experimental.
“Startups are special in their ability to throw stuff against a wall and try really hard things.”
This was Lindsay’s insight and it was the moment the room really leaned in.
If you’ve ever managed a thesis committee, you’ve already practiced stakeholder management. Think about it. You selected your committee strategically: someone for expertise, someone for connections, someone for the name on your CV. You figured out how often to update them. Not too much, not too little. You learned how to present honest progress when things weren’t going perfectly.
That translates almost directly to investor relations. Lindsay said:
“Don’t worry about disappointing them. If you’ve ever disappointed your thesis committee, you’ve already done it.”
She also talked about co-founder matching. A lot of MBA programs run speed-dating events where scientists meet business students. But early on, most investors actually expect the founders to be doing the business work themselves. Lindsay and her co-founder were both technical founders. He took the CEO persona, she took CTO. He read a couple of books and did the NSF I-Corps program, which is essentially a crash course in building a company from an idea.
Her point: you don’t need a business degree. You need to be willing to learn it while doing it.
“You will learn more practically by building something and making mistakes than theoretically in a classroom.”
Kyle works at the intersection of academia and commercialization, helping scientists navigate the gap between having an idea and actually building something. His most important message was about intellectual property. And he illustrated it with his own mistake.
During his PhD, he’d shown that targeting gut microbes could halve the dose of corticosteroids for treating inflammatory bowel disease. Really promising results.
But he’d presented the data at a conference two years before thinking about patents. That presentation became prior art. He couldn’t patent it.
He said it directly to the room: if you’re presenting at a conference like HUPO on technology you haven’t filed IP protection for, that’s now prior art. It will affect your ability to patent and it could undermine the core value of any company you build from it.
It’s boring. We acknowledged heads were nodding off. But it’s the kind of boring that can cost you everything.
His advice: if you’re at an institute and even vaguely thinking about spinning out, talk to your tech transfer office now. Not after you present. Not after you publish. Now. And get everything in writing. Parag reinforced this. When he had his Nautilus idea, he went straight to Stanford’s Office of Technology Licensing. They weren’t interested in filing, so they released it. But he got that in writing. If he hadn’t, Stanford could have come back years later claiming ownership.
Baljit’s journey was different from everyone else’s. She spent a decade at SCIEX, starting as a demo chemist. If you’ve ever been in the applications lab at a vendor, you know the role. Running samples, troubleshooting, talking to customers.
She moved into product management, which she described as “the ugly side of the business. You get to see every crack in the product.” But that experience gave her something most scientists don’t get: the full lifecycle from paper concept to commercialization. She saw what it takes to bring an instrument to market. The ZenoTOF launched a couple of years ago. She was on the project team in 2012. Ten years.
That’s what got her thinking about startups. How can you leapfrog that process?
What I loved about Baljit’s story was her honesty about the learning curve. She’s a scientist. She didn’t know what revenue, profit, or loss meant. A senior mentor at SCIEX picked her up early. Twenty-minute phone calls every couple of weeks. “Do you know what revenue means? Do you know what top line and bottom line mean?”
Eventually Baljit did an online MBA through Harvard Business School. She grew the SCIEX omics business from single to a triple-digit $M business across seven and a half years.
But her advice wasn’t “go get an MBA.” It was: find a mentor. Someone who’ll teach you the language. Someone who picks you up when you don’t even know what you don’t know. The formal education helps, but the relationship is what gets you through.
“Your sole responsibility as a founder is to kill ideas”
Baljit dropped this line and I watched the room flinch. Scientists are trained to persist. You run PCRs for 18 months on a single gene because that’s how a PhD works. You push through failed experiments because persistence is the whole game.
In a startup, that instinct can kill you.
“I did PCRs in my PhD for 18 months on a single gene. It sucked. I could not do that in biotech. I would surely lose all my investors.”
Your job isn’t to fall in love with one idea and make it work at all costs. Your job is to test fast, kill what doesn’t work, and find the thing that does. The founders who struggle most are the ones with brilliant technology and fifty potential applications. They try to do everything. The ones who succeed pick two or three areas, work through them one by one, and aren’t afraid to cross things off the list.
If you’re a scientist, that probably feels deeply uncomfortable. It felt uncomfortable to the people in that room. But multiple panelists said the same thing: the ability to let go of something you’ve invested time in is the single biggest mindset shift from academia to entrepreneurship.
Here’s something that came up throughout the panel but never as its own topic, and thinking back on the timing of that event, I now think it deserves specific attention given this new era we are in in 2026.
The speed at which you can build right now is fundamentally different from even months ago. The panelists had started their companies before the current wave of mature AI tools, programmable infrastructure, and open-source platforms that are available today. They had to build everything mostly from scratch. Many of the barriers they described, finding a commercial kitchen for chutney, renting lab time because you can’t afford your own, spending months setting up analytical pipelines, are being compressed.
That compression changes the math. The $100,000 it used to take to prove a concept might now be less time and money. The bioinformatics bottleneck that used to require hiring a specialist might now be something you can prototype on a well-structured platform or infrastructure. The ten-year timeline Baljit described at SCIEX? Startups are proving you can collapse parts of that dramatically when your infrastructure is designed for speed.
But here’s the tension. Moving fast is only valuable if you’re building on something solid. Speed without structure just means you fail faster and learn less. The founders on this panel who succeeded didn’t just move fast. They moved fast on top of a foundation: clear IP, the right mentors, structured thinking about where their technology fits. The ones who struggle are the ones who move fast in every direction at once.
For scientists thinking about this path today, the barrier to getting started is lower than it’s ever been. The tools exist. The platforms exist. The cost of testing an idea has dropped dramatically. What hasn’t changed is the need to be rigorous about what you’re building on top of. The infrastructure matters as much as the idea.
A postdoc asked the question everyone wanted to ask. Most of us don’t have Jeff Bezos’s dad handing us $50K. How do you get started when you need to house and feed your family?
Lindsay was candid. She started Talus in 2020 (best time ever to start a biotech company, she said, and those times aren’t coming back). She didn’t have a big-name PI behind her. She didn’t have family money.
Her practical advice:
Apply for SBIR grants. You’ll get rejected a lot. Apply for every accelerator you can find. You’ll get rejected there too. She called the whole thing “nice rejection.”
But: you only need one person to say yes. One check. Once one person is in, others get FOMO. The first investor de-risks the next one.
Kyle added that if you’re in the US, there are NIH mechanisms specifically designed to bridge from academia to entrepreneurship. A K99-style mechanism that funds the last years of your postdoc while you build foundational technology, then gives you additisonal runway in the startup phase.
Parag reframed the whole funding question. The key isn’t “I need X dollars.” It’s knowing exactly what you can prove with a small amount. “What is the thing you think you can do with $50,000?” That makes it infinitely easier for someone to write a check.
And I shared our story at Mass Dynamics. Four co-founders, which meant we could work for free in exchange for sweat equity. Our first grant was $20,000 and we thought we’d won the lottery.
One thing the panelists agreed on: accelerators are underrated. Y Combinator does life science. Creative Destruction Labs (CDL) runs programs specifically for computational health, ML, and deep science. Startmate in Australia. They don’t just give you money. They give you the network and the credibility that makes the next conversation easier.
Near the end, I said something that may have landed differently than the business advice. I talked about vulnerability.
“Scientists literally have a meeting called a defense,” I said. “You wear a lot of armor in what you do.”
My observation: one of the most powerful things I have learned in my career is that vulnerability is actually accepted in the startup world. You don’t need to know everything. The most powerful thing you can say is “I don’t know, but I’m going to find out.”
Parag took it further. When I asked for closing advice, he talked about ego and insecurity as the dominant failure mode he’s observed in founders.
“This experience has taught me to spend a lot of time staring at myself to uncover my insecurities that lead me to certain behavior. Spending that effort on introspection, not self-absorption, is probably the single most valuable thing I’ve done on this journey.”
That resonated with me. Because whether you’re founding a company, leading a lab, managing a core facility, or trying to get a difficult collaboration to work, the same thing is true. People can’t work together well when they’re operating from insecurity, from the need to be right, from armor they don’t even know they’re wearing.
The thing that struck me most wasn’t the business advice. It was how many people in that room were surprised to hear that the skills they already have, managing a thesis committee, navigating IP, articulating a research question, being resilient through years of failed experiments, are exactly the skills that translate.
Whether you ever start a company or not, the ability to clearly articulate what problem you’re solving, to be honest about what you don’t know, and to build trust with people who think differently than you: that’s not entrepreneurship. That’s science done well.
And the tools to act on it are better than they’ve ever been. The cost of testing an idea is dropping. The infrastructure to move from question to answer is getting faster. The founders on this panel built their companies in a world where most of that didn’t exist yet. You’re starting from a different place.
Watching a room full of scientists light up when they realized the gap between “where I am” and “where I’d need to be” is smaller than they thought? That’s the thing that keeps me doing these panels.
If you were in that room in Toronto, I hope this captures some of what made that session special. If you weren’t, I hope it gives you something useful.
And if any of this sparked something for you, come find us at the next one. Lindsay and I are doing it again at ASMS in San Diego on Wednesday June 3 5:45-7:00pm. See you then!
---------
This blog post was produced by Paula Burton using a combination of notes from presentations, discussions and insights. Final compilation was completed with assistance from ChatGPT 5.2. Each speaker reviewed their own section. No speaker reviewed the full piece end-to-end, and all synthesis and takeaways are my own. Any errors or omissions are unintentional, and the content is provided for informational purposes only. The views, thoughts, and opinions expressed in this text belong solely to the author, and not necessarily to the author's employers, organization, committees or other group or individual