Hertz Fellow Soyoun “Soy'” Choi, a graduate student at Stanford, is building machines that can sense and respond to their surroundings like humans. Already, her robotic dental assistant is working in clinics.
Every time a dentist treats a patient, they depend on a helping hand—an assistant who can remain still, holding a suction wand and retracting tissue. Vital to the field of dentistry, dental assistants require extensive training and the field faces chronic labor shortages and high burnout rates.
Soy Choi, a first-year doctoral student at Stanford University and 2025 Hertz Fellow, has developed an alternative to dental assistants: highly responsive robotic hands that can carry out the same repetitive tasks. Choi, who studies how to create robots with human-like dexterity and sensing capabilities, is the co-founder and CEO of Dentronic, the company behind Dexor, a robotic chairside assistant for dentistry. Dexor is designed to help dental teams with suction and soft-tissue retraction — essential tasks that every procedure depends on.
Dentronic’s first system, Dexor, is already working alongside dentists in clinical settings. In March, the company announced $1 million in pre-seed funding.
“Dentistry is a strangely perfect intersection of highly repetitive, physically demanding work and a desperate labor shortage,” Choi says. “Ultimately we are hoping to expand to other medical settings but we wanted to showcase that our robot can work safely in the clinic, and dentistry is an ideal place to test that.”
But medicine isn’t the only place that Choi wants to make an impact. She sees dexterous, AI-powered robots as tools to work alongside experts, helping to advance humanity in areas from climate change to natural disaster relief. For Choi, that potential comes with a sense of responsibility: robots operating in high-stakes environments must earn trust incrementally, which is exactly why Dentronic starts with repetitive, supervised tasks that support clinical teams rather than replace them.
From the orchestra to the lab
Choi’s path to robotics began in third grade, when she was struggling to learn the violin. Her father pulled up a video to show her: a humanoid robot, designed by Toyota, that could play a simple tune on a violin.
“It was really inspiring that a robot could play an instrument that was so difficult I could barely even make noises on it,” she recalls. “I became interested in how you can make a robot like that.”
Alongside motivating her to practice her violin, the video turned her interests toward engineering. By the time she arrived at Harvard as an undergraduate, Choi had already immersed herself in science both in school and through national competitions. She knew that she wanted to build robots that could sense and respond to the physical world the way that living things do.
At Harvard, Choi worked in the Microrobotics Lab of Robert Wood. Most robotic hands, she realized, were overly complex and not very human-like. They also couldn’t sense their own position without a camera.
“Humans combine vision, touch, proprioception, hearing and even subtle cues like skin deformation almost instantly,” Choi says. “Robotic systems just can’t do that yet.”
For her senior thesis, she designed a sensor system for robotic grippers that used stretchable materials to detect a finger’s position in real time, with no onboard vision needed. She also contributed to Project CETI, an initiative to decode sperm whale communication, through attachable robotic systems.
Deeper water
At Stanford’s Biomimetics and Dexterous Manipulation Lab (BDML), Choi is now extending that work into one of the most difficult environments on earth. Working with Mark Cutkosky and Oussama Khatib, she contributes to OceanOneK/OceanFive—a humanoid underwater robot being developed to operate at depths of up to 5,000 meters, where salt water and high pressure destroy most traditional sensors.
Dentistry and the deep ocean may seem like disparate applications of robotics, but for Choi, they both get at some of the same challenges. How can a robot sense and respond to its environment, even faced with ever-changing, moving conditions?
“It’s really new territory that we’re jumping into, but it’s very exciting,” Choi says.
In the Stanford pool with swimmers curiously looking on, Choi and her colleagues have been testing the newest underwater robotic sensors. Ultimately, they hope to design robots that can recover artifacts from shipwrecks, rescue divers or respond to natural disasters in environments too dangerous or remote for human crews.
It’s high-risk work that may not pay off immediately. That’s where Choi’s Hertz funding is game-changing, she says.
“This fellowship allows me very unusual flexibility to pursue ambitious, high-risk work across both research and entrepreneurship,” she says. “A lot of graduate students are really constrained by narrowly scoped funding sources and teaching obligations. Hertz gives me the room to work across areas and I’ve been very grateful for that.”
Two different worlds
Beyond the financial flexibility, Choi says, the Hertz community itself has shaped her trajectory. Through Hertz networking events, mentoring and partnerships that span research, industry and investment, she has connected with people who can engage seriously with both the science and the entrepreneurship side of her work.
“The people in the Hertz community are very action-oriented,” she says. “That directly helps me, both in my research direction and the questions I have about my startup.”
That balance—between her academic research and her entrepreneurship—is something that Choi, as a full-time graduate student with a startup on the side, is continually working on.
“I think that it is two different worlds. Academia generally rewards depth, rigor and long-term horizons, while startups are really about speed, execution and immediate usefulness,” says Choi.
But she also sees the value in working on both fronts at once. For her, both her academic work and her startup tie to her deep interest in robotic sensors. Each one builds off the other.
At Dentronic, Choi is balancing the speed of a startup (50 Dexors are already working in dental offices) with the long game required for AI-based learning. Over time, Dentronic hopes to use carefully collected workflow data to help Dexor learn from clinical environments and support more complex assistant tasks autonomously. “Just like a human surgeon doesn’t become a surgeon overnight,” Choi says, “our robot can start as an assistant, watch a lot of procedures, and train upon that.”
In that goal, Dexor isn’t so different from Choi—a first-year doctoral student already running a funded company, learning through experience, and getting better at everything she’s aiming to do.
About the Hertz Foundation
The Hertz Foundation is the nation’s preeminent nonprofit organization committed to advancing American scientific and technological leadership. For more than 60 years, it has stood as an unwavering pillar of independent support through the renowned Hertz Fellowship, cultivating a multidisciplinary network of innovators whose work has positively impacted millions of lives. Learn more at hertzfoundation.org.
