Inspiration
When I first registered, I started exploring different cybersecurity-related issues and read a few case studies to understand this area better. I came across this problem and felt fascinated, which led me towards building a solution for it
What it does
Vigilio is a platform that monitors sensitive data at runtime and learns normal behaviour. It tracks and analyses behaviour, and when the behaviour differs significantly, it identifies it as an anomaly.
How we built it
I built VIGILIO using Python and Flask for the backend, with HTML, CSS, and JavaScript for the dashboard. The system is divided into several parts: a runtime monitoring layer, a behavioral baseline, an anomaly detector, a remediation engine, and a security analyst component. The baseline is built from previously observed behavior. During a simulation, the system compares the current runtime behavior against that baseline and determines whether the behavior is normal or unusual, then connected everything to a web dashboard and deployed the application so the system could be accessed and demonstrated through a live website.
Challenges we ran into
One of our biggest challenges was making the concept work as an actual runtime flow rather than just a theoretical security model. I had to make sure that the monitoring, baseline, anomaly detection, and remediation components communicated correctly with each other. Deployment also brought its own challenges. Things that worked locally, especially around file paths, baseline loading, and environment configuration, needed to be adjusted for a production environment. Time was also spent in refining the dashboard so that the security information was easy to understand without making the interface feel overloaded.
Accomplishments that we're proud of
I am proud that I was able to take the idea from a security concept and turn it into a working end-to-end prototype. VIGILIO doesn't just detect unusual behavior. It connects monitoring, detection, analysis, and automated remediation into one flow. Also, proud of the dashboard because it gives a clear visual representation of something that would otherwise happen silently in the backend. Most importantly, we now have a live, deployable version of the project that we can actually demonstrate rather than just explain.
What we learned
Building a security system is not only about detecting threats. Context and behavior matter just as much. Also learning about designing a system where multiple components have to work together reliably, from runtime monitoring to anomaly detection and remediation.
What's next for VIGILIO
next goal is to make VIGILIO more capable and closer to a real-world runtime security layer, to expand monitoring beyond the current sensitive-resource simulation, support more types of sensitive data, collect richer runtime behavior, and improve the anomaly detection model. The current version is our proof of concept. The larger vision is to make VIGILIO a system that can continuously understand how sensitive data behaves inside an application and respond when that behavior becomes risky.
Log in or sign up for Devpost to join the conversation.