Two Manassas IT Students Build a Bot Detector and Warehouse Inventory System
Two students at ECPI University’s Manassas campus turned the practical skills they learned in class into technology solutions that went well beyond the classroom.
Timothy M. Reeser built a detection tool that catches bots pretending to be real people, and Justin Robinson built an inventory system for a local wholesale carpet and flooring distributor.
How do you catch a bot that acts like a person?
It starts with identifying what normal is, then flagging anything that drifts away from it. That’s the approach behind BSV-RBAM (Behavioral Synthetic Validation via Recursive Baseline Anomaly Matrix), the forensic analysis engine that Timothy built.
Timothy completed his Bachelor’s in Information and Cybersecurity Operations at ECPI University and is set to begin his master’s program this October.
He spent the past year building the engine in Python and stress-tested it on data sets ranging from 50 to 500,000 data points, including a coordinated 10,000-unit bot army. A stabilization metric he developed, the Reeser Coefficient, locks the baseline in place in real time, so a threat can’t slowly nudge it without getting caught.
When the engine flags something, it gives SOC analysts a written report and a heatmap showing what changed. Timothy has also filed a provisional patent with the United States Patent and Trademark Office (USPTO) for this technology.
“He didn’t try to build a better alarm,” says Dr. Vicente Smith, Associate Program Director of Information Technology and Cyber Security. “He built something that understands what normal is supposed to feel like. And he’s building it while still finishing his own education.”
How do you keep a whole warehouse working from the same inventory data?
Everything passes through one central application, and the warehouse machines never touch the database directly. Justin, a Computer and Information Science student, built a system around this design for the distributor, which was tracking inventory by hand.
The company needed several people working from the same live data without paying for a full database server. Justin took on the project alone, serving as both developer and IT consultant. His solution began as a basic if/else script for a Python class assignment.
The system runs as two connected applications sharing one database: a desktop app for the office and a web app that anyone on the warehouse floor can open in a browser with nothing to install. Every request passes through the central application, which helps prevent corrupted data and conflicting writes.
Justin also designed and set up the network the system runs on, with routing, addressing, and firewall rules configured to keep the server off the open internet. The project took about four months and involved on-site troubleshooting alongside the company to figure out what the business actually needed.
“Justin built exactly what this business needed and nothing more, and knowing where to stop is a harder skill to teach than any programming language,” says Dr. Smith.
