Luyen Chou spent the beginning of his career trying not to become an educator. The family precedent was almost comically thorough: both parents, plus his grandparents, had worked in education. As a boy he cared about marine biology and astronomy, and he worried that teaching would make those interests mutually exclusive. Then his old high school in New York asked him to come back after college and teach history. Chou expected a short detour. The detour became the road.
He fell for the classroom, and for the freedom hidden inside it. Teaching did not require him to surrender his other curiosities. It gave him an audience for them. In the late 1980s he was already bringing computer simulations and digital images of archaeology into ancient-history lessons at the Dalton School. The hardware was primitive. The instinct was not: let a new tool make an old subject more available, more visible and more alive.
That instinct now sits inside DeweyLearn, the company Chou co-founded with applied-AI specialist Dirk Liebich. Its software analyzes video, audio and learning data to evaluate people doing things: a culinary student handling a knife, a trainee practicing a professional skill, an instructor guiding a room. Instead of merely scoring a final answer, the system is designed to notice the performance that produced it.
A computer in one column, a classroom in the other
Chou's first computer was a TRS-80. At Harvard, where he studied philosophy, computing was a hobby and a way to earn money for gear. Education occupied another intellectual column: history, philosophy and the philosophy of education. He disliked much of the educational software around him because it narrowed thought rather than opening it. For a while, technology was technology and education was education.
The columns eventually met. In 1992 he founded Learn Technologies Interactive, an educational-software business whose multimedia work included Qin: Tomb of the Middle Kingdom. The 1996 title turned imperial Chinese history into an exploratory game; Chou was credited as a program designer and executive producer. This was the CD-ROM age, when a shiny disc could be treated as a small museum, textbook and theater at once. Chou's business was already testing the question he would keep asking through several technical eras: what is the machine good for, educationally?
His next laboratories were literal schools. Chou helped design and launch The School at Columbia University in 2003, serving as associate head and leading its Center for Integrated Learning and Teaching. At Teachers College he met educator Daniel Rubenstein. Together they co-founded Brooklyn Prospect Charter School, which opened in 2009 with an International Baccalaureate curriculum and a plan to grow one grade at a time. Building a school differs from shipping software in one obvious respect: the users walk into the room every morning. The feedback is immediate, human and difficult to reduce to a dashboard.
The product years
Schoolnet brought Chou into the era of large-scale educational data. As chief product officer, he worked on curriculum, assessment and reporting tools for districts and states. Pearson acquired Schoolnet in 2011. Chou spent seven years at Pearson and eventually became chief product strategy officer, responsible for strategy across a global product portfolio then described as worth $6 billion. It was an education in scale: many countries, many institutions, and no single definition of a learner's need.
In 2018 he joined Trilogy Education as its first chief product officer, overseeing academics, product management and technology for programs that taught working adults coding, analytics, design and cybersecurity in partnership with universities. When 2U acquired Trilogy the following year, Chou moved into the combined company and later became chief learning officer. His remit sounded simple and was not: define and protect the quality of learning experiences across products.
During the sudden migration to remote instruction in 2020, his team created “Studio in a Box,” a portable production setup that let university faculty record professional lessons away from campus studios. The object was practical, almost domestic: equipment delivered to the teacher, not the teacher delivered to the equipment. It expressed a Chou habit. Technology should travel toward the people whose expertise gives it meaning.
Outside those operating roles, Chou kept one foot in the institutional questions around technology. In 2024 he joined Learning.com's board while advising education organizations on product strategy and artificial intelligence. The board appointment returned him to K-12 digital readiness, including the less glamorous work of helping students and teachers use technology safely and responsibly. New tools, in his telling, arrive with both potential and obligation.
The observation problem
Generative AI made education software conversational. A student could ask; a machine could answer. Chou saw another possibility. Cameras and microphones capture physical action, timing, tone and interaction. A model that can work across those modes might help evaluate not just what somebody knows, but what that person can do.
DeweyLearn builds its assessments around each institution's own curriculum, standards and expert judgment. The system documents evidence from observed behavior and returns feedback against those frameworks. Its operating principles are pointedly unromantic: private by construction, grounded in evidence, custom by design and expert in the loop. The machine supplements judgment. It does not get promoted above it.
The feedback loop
That approach has moved beyond the conventional classroom. By July 2026, more than 2,000 culinary students were receiving targeted feedback through the platform. Other reported uses included professional training and corporate competency assessment. The variety matters because Chou's argument is larger than school software. Learning happens anywhere a person practices, receives correction and tries again.
The company won the 2026 ASU+GSV Cup after being selected from more than 3,000 nominees. In July it announced an oversubscribed $5 million Series A led by SJF Ventures, joined by Catalysis Capital, Morningside and Owl Ventures. Funding is not a theory of learning, but it gives the theory room to be tested in more places.
A name with a long memory
DeweyLearn's name is a design principle wrapped in family history. The philosopher John Dewey argued that people learn through doing. During his travels in China between 1919 and 1921, he encouraged Chou's grandmother to study teaching at the University of Chicago. A century later, her grandson attached Dewey's name to a system intended to evaluate learning as performance rather than recitation.
There is another inheritance nearby. Chou's father, Chou Wen-chung, was a Chinese American composer and Columbia professor known for joining Chinese and Western musical languages without dissolving their differences. Luyen and his brother Sumin have helped tend that legacy, representing their father at the 2018 opening of the Chou Wen-chung Music Research Center in Guangzhou and participating in centennial programs after his death.
The parallel should not be pushed too neatly, but it is hard to miss. One generation worked between musical traditions. The next works between teacher and machine, trying to keep the character of each intact. Synthesis is valuable only when the ingredients remain legible.
What the builder kept
Chou has offered aspiring education entrepreneurs advice with the briskness of someone who has watched several waves arrive and recede: passion first, then the skills needed to serve it. He has also said that every failure is worth ten successes. It is a teacher's ratio. Success confirms one route; failure opens ten questions about why the learner, lesson, company or machine behaved as it did.
His own career reads less like a ladder than a repeating lesson plan. Enter a room. Notice what people are trying to learn. Build something. Watch what happens. Revise. The tools move from a TRS-80 to a CD-ROM, from a district dashboard to an online studio, from a chatbot era to multimodal models. Attention is the durable technology.
DeweyLearn is Chou's latest attempt to scale that attention. Its promise is not that a machine becomes a master. It is that more learners can feel the useful pressure of a master's eye while the real experts keep writing the standards. For someone who once thought teaching might crowd out his other passions, Chou has built a life around the opposite discovery: teaching is where the interests meet.