When Mark Hallen, ’06, arrived at Cary Academy as a self-described “bookish” student with a budding interest in math and science, he could never have imagined that one day he’d help turn what once felt like science fiction into reality. Today, he is doing just that, leading a team of scientists in designing cancer drugs with the aid of AI.
As co-founder and Chief Technology Officer of Ten63 Therapeutics, a Durham-based biotech startup, Hallen is pioneering a radically new interdisciplinary scientific approach—one at the cutting edge of laboratory science, computational molecular modeling, and artificial intelligence—to create effective therapies that target some of the most stubborn and difficult-to-treat cancers.
His journey is one of purposeful exploration that took root in his high school chemistry labs and classrooms and has evolved into a career focused on saving lives.
The catalyst
Hallen traces his professional beginnings directly to Cary Academy. “CA is really where I started getting into what I’m working on now—especially on the chemistry side,” he reflects. A student who thrived in both science and math, he found particular inspiration in the challenges of Mr. Gray Rushin’s chemistry class.
But it was a history course that hinted at something more profound, a reminder that science didn’t exist in a vacuum. “It offered a different kind of lens—how systems impact people,” Hallen recalls. That early insight stuck, shaping the kind of problems he seeks to solve in his career: significant, real-world challenges with human consequences.
It wasn’t just CA’s academics that proved formative. Running cross-country and track offered essential life lessons—grit, persistence, and a team-oriented mindset—that he draws on today. “Research can be like a race where you’re all training together and helping each other improve, but at the end of the day, everyone still has to do their part, to run their own race,” he says.
After CA, Hallen headed to Duke University, double majoring in chemistry and math. He quickly found himself drawn to interdisciplinary work, attracted to problems that sat at the intersection of biology, computation, and medicine. It was during an early research stint in a cell biology lab that he discovered a surprising strength, not at the bench, but in computational analytics.
“At the time, we were studying fruit flies, using microscopy to study cell division and gene expression. I started building algorithms to analyze the data we were collecting,” shares Hallen. “That was a turning point; I realized I was contributing more with my digital analysis than the benchwork,” he explains. That shift—from hands-on experiments to algorithmic coding and computational modeling—would define the next decade of his research.
From microscope To molecule
At Duke, Hallen became fascinated by the potential of computers to model living systems. While “computational biology” can mean many things, for him it came down to this: using math and code to understand how biological systems work—and how to fix them when they don’t.
He was especially drawn to proteins—the tiny building blocks of life involved in nearly all cellular processes. In high school, the idea of designing them on a computer felt like science fiction. But by his senior year at Duke, Hallen learned of the Donald Lab, where that fiction was fast becoming reality. Led by Dr. Bruce Donald, James B. Duke Professor of Computer Science, Mathematics, Chemistry, and Biochemistry at Duke University, the lab had just published research showing how computational tools could be used to design and reshape proteins for entirely new functions.
“It was the first time I saw that you could really plan molecular behavior on a computer with a high degree of accuracy,” Hallen recalls.
He joined the lab for his PhD, delving deep into algorithmic modeling and developing methods to search efficiently through an almost unfathomable universe of possible molecules to identify those most likely to succeed as drugs.
How vast is that search space? “There are about 10⁶³ possible drug-like molecules,” Hallen explains. “That’s one followed by 63 zeros. We can’t begin to test all of them in a lab—or even simulate them one at a time on a computer—to find the one that might be a useful drug for a specific application.”
In the face of that challenge, Hallen set out to design algorithms to search intelligently through that vast chemical space, identifying the best candidates with the highest mathematical guarantee of success. In theory, it’s a solution that shares a similar logic to internet search engines: not perfect, but fast, innovative, and shockingly effective.
Algorithm to enterprise
That insight—applying sophisticated search algorithms to molecular design—became the foundation of what Hallen would later call molecular voxel theory (MVT). This algorithmic framework breaks molecular movement into multi-dimensional units (the higher-dimensional equivalent of pixels) that can be modeled and searched with remarkable precision, enabling high-resolution drug design.
A subsequent role as a research assistant professor at the Toyota Technological Institute in Chicago gave Hallen broad latitude to develop new directions for molecular voxel theory. Some of these directions could have a high impact on the design of new medicines, but would require experiments beyond the budget available in that role. Then came a call from a former labmate: “I think we’re ready to start a company.”
The timing was right. The mission was clear. In 2018, they launched Ten63 Therapeutics—the name a nod to the staggering molecular space looming before them—with Hallen leading research and co-founder Marcel Frenkel handling the business side. His former mentor, Dr. Bruce Donald, chairs the company’s scientific advisory board.
Cracking the undruggable
Ten63’s first challenge was an ambitious one: to design a molecule that inhibits Myc, a protein overexpressed in many aggressive and “undruggable,” or difficult-to-treat, cancers. Myc isn’t mutated in most cases; it’s just turned up too high. “It interacts with DNA and speeds up cell division,” Hallen explains.
Ten63’s solution? Combine Hallen’s algorithms with advanced machine learning through their proprietary BEYOND platform, an AI framework that explores vast chemical spaces with greater efficiency and precision than traditional trial-and-error methods. Trained on simulations grounded in first-principles physics and refined by real-world experimental feedback, BEYOND doesn’t just predict what might work; it reveals why a molecule works and how it can be optimized to meet specific therapeutic goals.
That insight is essential in designing a Myc inhibitor, which requires modeling how candidate molecules and the Myc protein interact, flex, and twist in 3D space. From there, molecules are refined, synthesized, and tested through increasingly complex rounds of experimentation to determine whether they behave as intended, from cell to mouse and, eventually, to human trials.
It’s painstaking, high-stakes work, but rewarding in its promise. “It’s a lot of iteration,” Hallen says. “But we just had our first experiments show that one of our compounds slowed tumor growth in mice.”
Discovery to impact
Ten63 aims to advance its lead compound into human trials by 2027. “That would be a huge milestone,” says Hallen. But for him, success has always been about something more. “Improving outcomes for patients—that’s the real success.”
Hallen remains energized by the challenges ahead, bringing together data, domain expertise, and determination to unlock new frontiers in medicine. But he’s also thoughtful—and clear-eyed—about the rapidly evolving landscape of biomedical research. Artificial intelligence, he says, has changed everything.
“There’s been so much hype that it’s crowded out other research directions and methodologies,” he reflects. “But not everything needs to be solved by AI alone. Some of the most promising advances in our field come from combining machine learning with deep domain knowledge and experimental innovation.”
According to Hallen, the most effective work lives at the intersection of AI and hands-on science. “That’s a hard part of the work—being in a position to straddle both domains and know when you’re making mistakes in either one,” he says. If he didn’t understand how molecules behave in the real world, he explains, he’d be designing things that chemists might immediately reject. And if he didn’t understand the modeling, he wouldn’t know how to explore the search space effectively.
That deep, dual fluency—in both molecule and model, code and machine learning—isn’t just a technical advantage. It’s a reflection of a broader ethos that defines Hallen’s work: a belief that transformative progress happens at the intersection of disciplines, where science, technology, and humanity converge.
It’s an ethos rooted in his earliest experiences at Cary Academy, where a passion for chemistry and math met a systems-level view of the world. Today, it drives his work at the forefront of biotech innovation—a path defined not just by technical breakthroughs, but by a commitment to purpose and impact. With each algorithm and experiment, Hallen isn’t just imagining the future of medicine—he’s helping to build it, one molecule at a time.