Homagama, Sri Lanka · SLIIT, Year 3
Cybersecurity undergraduate working across network security, penetration testing and secure application development.
Ranked top 7% globally on TryHackMe. I build full‑stack and AI‑integrated systems alongside my security work — most recently an AI shopping agent recognized in a national developer competition.
I'm a cybersecurity undergraduate at the Sri Lanka Institute of Information Technology, specializing in network security, penetration testing and secure application development. Outside coursework I spend most of my time in TryHackMe rooms and building full-stack projects — usually the two intersect, since I like understanding a system well enough to secure it, which means building one first.
Recent work includes an AI-integrated shopping agent built on the Model Context Protocol, a role-based customer care platform on Spring Boot, and a segmented enterprise network design for 75+ devices across six departments. I'm currently deepening my research side too, with a paper tracing IoT security's evolution from the Mirai botnet to today's AI-driven threat landscape.
Grouped the way I'd actually list them on a machine, not scored out of 100 — a stack either gets used or it doesn't.
Selected builds spanning AI integration, enterprise networking, and full-stack systems.
A multi-modal shopping assistant integrating Gemini 2.5 Flash via the Model Context Protocol, for conversational product search and comparison.
Recognized among 700+ national entrants — Certificate of Participation (Builder track), Kapruka Holdings PLC, July 2026.
Full-stack support platform for 6 user roles on a layered Spring MVC → Service → JPA architecture, covering ticketing, repairs and inventory. Automated stock monitoring and reordering thresholds to cut manual purchase-order tracking.
Secure 3-floor network for 75+ workstations, servers and IoT devices, using VLAN segmentation and /27 subnetting across 6 departments. Hardened wireless and remote access with WPA2/WPA3, VPN configuration and centralized firewalls.
AI web app classifying 120+ dog breeds via a gatekeeper filter plus a ResNet50V2 classifier. Improved prediction accuracy using transfer learning, data augmentation and test-time augmentation on the Stanford Dogs dataset.
Arduino-based smart home system for remote and voice control with automated environmental monitoring, and real-time control logic that triggers device responses from live sensor data.
IE2092: Introduction to Cybersecurity — SLIIT, Year 2 Semester 2, 2026
Traced IoT security's evolution from pre-2016 negligence through the Mirai botnet watershed — over 600,000 devices and a 1.2 Tbps DDoS attack — to today's expanded attack surface.
Analyzed five emerging threats using peer-reviewed and government sources, including NIST and Veracode:
TryHackMe
Cisco Networking Academy
Kapruka Holdings PLC — one of 700+ developers across Sri Lanka to submit a working agent on the Kapruka MCP.
TryHackMe — AI Security and Cyber Security 101 in progress
Google Cloud Skills Boost (2025-CH2) · Networking Basics in progress, Cisco Networking Academy
Open to internships, security research collaborations, and interesting build projects.