An 18-year-old from New Jersey has built an AI model that predicts how individual tumors will respond to chemotherapy, using genomic data to help doctors select optimal cancer treatments and avoid unnecessary side effects. Mythreya Dharani's project placed him among the top finalists in the Regeneron Science Talent Search, the Society for Science's prestigious high school competition that draws roughly 1,900 entries each year.
The model analyzes genomic information to forecast treatment outcomes at the patient level, a shift from the one-size-fits-all approach that often governs oncology protocols. For clinicians, the practical stakes are direct: better predictions mean fewer rounds of ineffective treatment and a clearer path to therapies that actually work for a specific tumor profile.
Research beyond the lab
Dharani's work extends beyond the model itself. He is co-captain of his school's varsity debate team and volunteers as an instructor with the UCLA Math Circle, where he teaches college-level mathematics to elementary and middle school students.
The Regeneron Science Talent Search is unique among high school competitions in that it focuses on identifying and engaging the nation's most promising young scientists. Students submit original research across critical scientific fields, and the competition culminates in a public exhibition and an awards ceremony where the top ten winners are announced.
For those working in research roles, the project is a practical example of how AI for Science & Research can be applied to clinical decision-making - not as a theoretical exercise, but as a tool that addresses a concrete bottleneck in cancer care. The approach mirrors what professional research teams are increasingly expected to build: models trained on domain-specific data that produce actionable outputs for practitioners.
A separate milestone in astronomy
In a separate development, Atharv Sood, an 11-year-old student from India, has been credited with three preliminary asteroid detections through systematic observation and analysis as part of an international astronomical collaboration. The detections contribute to broader efforts to track near-Earth objects.
Sood's work highlights how access to the right tools and mentorship enables young researchers to make real contributions, even outside formal institutional settings. The astronomical community has taken note, and his efforts may encourage other students to engage with citizen science projects.
Why this matters for science and research professionals
Two patterns in these stories are worth attention. First, the chemotherapy prediction model shows that AI applied to genomic data can move from research paper to clinical relevance quickly when the problem is well-defined. Second, both projects demonstrate that meaningful scientific output no longer requires a formal lab appointment - a point that matters for professionals building research pipelines that include external collaborators or citizen scientists.
For researchers considering how to structure their own work, the takeaway is practical: AI Learning Path for Research Scientists resources now cover exactly these skills - from data preparation to model validation - that enable domain experts to build predictive tools without waiting for a dedicated machine learning team.
Your membership also unlocks:
