AlziDetect, a healthcare technology startup founded by student researcher Ryan Borui Zhu alongside Ali-Mansur Valiyev and Harihar Rengan, has launched a pioneering camera-based screening platform built to catch early signs of cognitive decline linked to Alzheimer’s disease. The technology could reshape how families and healthcare providers approach early detection, cutting through the cost, access, and wait-time barriers that have delayed diagnosis for far too long.
Right now, anyone worried about their cognitive health faces a long road. Specialist referrals, waiting lists that stretch for months, and imaging procedures that cost thousands of dollars all add up to a bottleneck that can push back intervention by years. AlziDetect’s platform aims to shortcut that entire process.
How It Works
Using a standard laptop camera, the platform maps 468 facial and iris landmarks in real time through Google’s MediaPipe framework. From this, it builds a picture of a person’s gaze trajectory, fixation stability, and eye movement velocity; these are patterns tied to early cognitive change in published neuroscience research. A full screening takes under three minutes and produces results at home, with no specialist visit, clinic booking, or dedicated hardware required.
“We’re bridging a massive gap between what research has proven possible and what patients can actually access,” said Ryan Zhu, who shadowed at a public hospital for three months before founding AlziDetect. “I’ve seen how early intervention in cognitive decline can change the entire trajectory of someone’s quality of life. Our mission is to make sure that financial constraints or geographical location never stand between a family and an answer.”
A Rigorous Scientific Foundation
AlziDetect’s methodology draws on peer-reviewed research and is backed by an active line of clinical validation work. Ryan Zhu, together with Ali-Mansur Valiyev and Harihar Rengan, authored a paper on the underlying science titled “Normative Bayesian EEG Analysis with Directional Priors for Interpretable Alzheimer’s Risk Screening,” developed with mentors at The Hong Kong Polytechnic University and published at IEEE COMPSAC 2026.
The research tackles a problem that has held back Alzheimer’s screening tools for years: most models need real patient data on which to train, and this data is expensive, hard to access, and ethically fraught to collect. AlziDetect’s approach circumvents this entirely. It learns what healthy brain activity looks like, then flags deviations using seven directional markers pulled from published Alzheimer’s biomarker research, without training on a single set of Alzheimer’s patient data. In early validation testing, the model distinguished healthy patterns from Alzheimer’s-consistent ones with an AUC of up to 0.938, a standard measure of diagnostic strength, and it performed just as well on the 19-channel EEG setups already standard in clinics today. Every score the model produces can also be broken down into the exact brain regions and frequency bands driving it, giving clinicians a transparent readout instead of an unexplainable number.
Alongside this, the team is building partnerships with clinical and hospital networks to validate the platform’s results against established diagnostic standards.
The project has also picked up recognition from several innovation programs, including a spot in the Blue Ocean Top 100, a Conrad Challenge Silver Award, and a Diamond Challenge Semifinalist placement.
Building the Full Ecosystem
Screening is just the starting point. AlziDetect is developing a broader platform that will include clinician-guided education to help users make sense of their results, tools to track cognitive changes over time, and features built for families to support long-term memory care.
A Young Team Driving Innovation
What makes AlziDetect stand out is the mix of technical depth and fresh perspective driving its team. Founders Ryan Zhu, Ali-Mansur Valiyev, and Harihar Rengan lead the research behind the screening approach, serving as first authors of the team’s Bayesian EEG paper and steering the platform’s clinical validation strategy in partnership with mentors at The Hong Kong Polytechnic University’s Departments of Computing and Optometry.
Chief Financial Officer and Head of Outreach, Kubby Bhatia, connects the technology to the communities it’s meant to serve, drawing on a background in AI, entrepreneurship, and economics to push for equitable access and real-world adoption. Eklavya Tomar, working as an AI and Health Integrated Engineer, builds the technical systems that sit at the intersection of AI and clinical healthcare needs. Neil Sharma, as a Healthcare AI Programmer, is developing the infrastructure that keeps the platform running.
Together, this global team of high school students is tackling one of healthcare’s toughest problems: turning years of laboratory research into something people can actually use.
What’s Next
AlziDetect is now in its clinical validation phase, working closely with healthcare partners to make sure the technology holds up to rigorous scientific standards before a wider rollout. The company is looking for clinicians, researchers, and healthcare systems interested in joining upcoming validation studies.
For more information, visit alzidetect.com.
About AlziDetect
AlziDetect is a healthcare technology startup building accessible, camera-based tools for early Alzheimer’s risk detection. Founded by a team of student researchers and entrepreneurs, the company’s goal is to remove the barriers standing between people and early cognitive assessment, so intervention can happen sooner.
