My grandfather clicked
a link. I built the fix.
It was a normal afternoon. My grandfather received a text message claiming his bank account would be suspended unless he clicked the link. He clicked it — not carelessly, but the way anyone would who doesn't know what a phishing link looks like. He entered his credentials. They didn't work. By the time he understood what had happened, the damage was done. And beyond the financial loss, something else broke: his confidence in using the internet at all.
That moment made something clear. If it had been my father — someone tech-savvy, someone who knows what a suspicious URL looks like — he wouldn't have clicked. But my grandfather isn't less intelligent. He's simply less familiar with a world that moved almost entirely online while many people his age were never given the tools to navigate it safely. Banking, groceries, appointments, communication — all digital now. And the people least equipped to handle the threats that come with that shift are the ones who need it to work most reliably.
That observation became the design brief for PhishSentinel: not a tool for IT professionals, but a tool for people like my grandfather.
total
& blocked
across all scans
I built a website where you paste any suspicious link and instantly find out if it's legitimate — using a machine learning model trained on phishing and legitimate URL patterns. I showed it to my grandfather. Walked him through it. Watched his confidence return. He told his friends. A week later he called me to his room where he was sitting with a group of them — and asked me to teach them too. His friends were less familiar with their phones than he was. I sat with them, one by one, until each person understood. Then I built the Chrome extension and got them all to install it — so the protection was automatic, invisible until needed.
Paste a link. Know in seconds.
Any user — regardless of technical background — can paste a suspicious URL and receive an instant, plain-language verdict. No jargon, no interpretation required. Powered by an ML model trained on phishing URL patterns.
Automatic protection while browsing
The Chrome extension uses the same backend to actively monitor current pages and flag or block suspicious sites in real time. The protection is invisible until it's needed. Installed on the computers of my grandfather's friends after building it.
Security education module
Building an interactive module that teaches users to recognise phishing indicators themselves — moving from tool-assisted protection to genuine digital literacy. Designed for the same audience: people who were never taught how any of this works.
Adaptive ML detection
Expanding the training dataset, improving detection accuracy for new phishing patterns, and exploring adaptive models that improve as the threat landscape changes — because phishing attacks are not static.
🛡️ Xavier Uni Cybersecurity Challenge — Team SecuroSphere
Separately from PhishSentinel, in Grade 11 I competed in the Xavier University Cybersecurity Challenge — Invisible Wars as part of Team SecuroSphere (with Yatharth Pant and Abeer Verma). A full scenario-based national security competition — threat assessment, incident response lifecycle, ethical decision-making, and long-term system hardening. Full details on Competitions →