High school is no longer the ceiling: a self-taught teenager trains AI against head and neck cancer at Penn State

By Penn State · July 17, 2026.

The facts are modest and well documented: Shaunak "Shaun" Dalal, a recent graduate of Hershey High School (Pennsylvania), has become the youngest member of a research team funded by the Penn State Clinical and Translational Science Institute (CTSI) that applies artificial intelligence to the diagnosis and treatment of head and neck cancers. Dalal discovered programming in ninth grade, learned Python by watching YouTube, completed online machine learning courses and, on his own, had already built a project to identify brain tumors in MRI scans. With a family history of cancer, in his senior year of high school he used a school internship program to reach out directly to Neerav Goyal, a head and neck oncologic surgeon and chief of that division at Penn State Health Milton S. Hershey Medical Center, and got him to accept him as a mentor. Beyond programming, he did clinical rounds with Goyal so as not to lose sight of the fact that behind every data point there is a patient.

The article gives no technical details —we do not know what model the team is training, on what data, or with what results— and it would be a mistake to inflate that part with conjecture: this is a story of a personal trajectory, not a paper with metrics. But the relevant fact is not in the algorithm, it is in the barrier to entry. A decade ago, joining a real medical AI project required at a minimum a doctorate in progress and years of formal training. Today, an introductory high school course, a YouTube channel and a handful of online courses are enough to acquire the technical foundation; what was missing —access to a clinical mentor willing to open the door— was here solved by the personal initiative of a high school student.

This fits a thesis we have been repeating from another angle, that of employment: competitive advantage no longer resides so much in knowing how to program as in knowing what to apply it to and with whom. Dalal stands out not for writing better code than a senior researcher, but for combining a personal motivation —illness in his family—, self-teaching accelerated by free tools and the ability to ask for an opportunity that was previously not even considered. It is the same dynamic we will see multiply: AI brutally cheapens the technical cost of entry, and what begins to grow scarce is access to mentors, clinical data and institutional trust. That access, today, remains very unevenly distributed, and it is worth saying so plainly: for every Dalal who finds a Goyal willing to open the door, there are many equally capable teenagers without that contact.

If we connect this to the underlying horizon we defend at Zendoric —an AI that helps bring us closer to eradicating diseases and prolonging health—, stories like this are not the flashy headline of a great breakthrough, but the quietest and most reliable signal: the army of people capable of working toward that goal is broadening from below, not only from above. The question that remains open, and that will probably define how long that horizon takes to arrive, is whether health and educational institutions will build systematic channels to capture that emerging talent, instead of relying on each Dalal having the initiative to write the right email.