Dilsha Mihiranga

PowerLens — Transformer Maintenance Record Keeper

Jul 2025 — Nov 2025

A comprehensive full-stack, AI-powered system designed to automate transformer maintenance by using computer vision to detect thermal anomalies and providing a digital platform for managing detailed inspection records and analysis.

My Contribution: Developed both the frontend and backend maintenance inspection module and integrated Python-based AI inference for thermal analysis. Built advanced React frontend tools for image annotation and designed the backend architecture to persist inspection data and AI detection results.

Tech Stack: Java Spring Boot, React (TypeScript), MySQL, Python, YOLOv11

RideWay — Bus Booking Platform

Dec 2025 — Present

Full-stack bus booking platform with a Node.js/Express REST API and a React + TypeScript frontend, providing seat selection, real-time trip management, and secure user onboarding.

Implemented JWT authentication with role-based access (admin/user), email verification via Resend, and MySQL + Sequelize models for Users, Buses, Trips, and Bookings. Deployed as Dockerized services on a DigitalOcean droplet with a custom domain.

Tech Stack: Node.js, Express, React, TypeScript, MySQL, Sequelize, JWT, Resend API, Docker, DigitalOcean

Demo Access: Customer — customer@test.com / customer123, Admin — admin@test.com / admin123 (or create your own account with email verification).

RFID-Based Inventory Management System

Jul 2024 — Dec 2024

Real-time inventory tracking system integrating a Spring Boot backend with a custom ESP32 RFID reader for automated stock monitoring, and a React.js frontend with WebSocket integration for live updates and MQTT for device communication.

My Contribution: Implemented the WebSocket/MQTT communication layer, designed the real-time inventory dashboard, and contributed backend enhancements.

Tech Stack: Spring Boot, React.js, Arduino/ESP32, RFID, MQTT, WebSocket

Multi Hop LoRa for Post Disaster Communication (Final Year Project)

Jun 2025 — Present

Engineered a decentralized multi-hop LoRa network for post-disaster rescue operations, simulating and analyzing routing protocols using OMNET++ to optimize network performance.

Developed a reliable large-data fragmentation algorithm for Structural Health Monitoring (SHM) and built an offline GPS tracking system using ESP32 hardware.

My Contribution: Created web and mobile interfaces for the offline GPS tracking system and contributed to algorithm development.

Tech Stack: LoRa, OMNET++, C++, ESP32 (Embedded C), Python, HTML, CSS, React Native

Asset Registry and Tracking System

Oct 2025 — Dec 2025

Microservices-based application for managing company assets, designed as part of the WSO2 Linux and DevOps Training program and deployed on a zero-trust Kubernetes infrastructure with automated CI/CD pipelines.

My Contribution: Developed a custom .NET Load Balancer service to efficiently manage and distribute traffic across the microservices architecture.

Tech Stack: .NET (C#), Kubernetes, Docker, MySQL, CI/CD

Techtro — Full Stack Web Application (Individual)

Jul 2024

Secure MERN stack application with JWT authentication and email verification, built to handle concurrent user requests with role-based access control.

Tech Stack: React.js, Node.js, Express, MongoDB, JWT

Vision Based Label Placement Verification System

Jan 2024 — Jul 2024

Vision-based system to detect misprinted labels on a production line with an integrated feedback loop, including electronic and mechanical systems to automatically reject faulty products.

Tech Stack: Python, YOLOv8, C++, Microchip Studio, Altium, SolidWorks

ClariFi — Medical Report/Scan Analyzing Assistant AI

Jul 2024

AI-powered assistant to analyze medical reports and scans, providing insights and automated suggestions to users through a mobile-first experience.

Tech Stack: JavaScript, React Native, Python, Django, Llama, Firebase

SDN Network Optimization & Traffic Engineering

Aug 2024

Comprehensive study on SDN-based network optimization by implementing Traffic Engineering and Quality of Service algorithms in a virtualized Linux environment using Mininet and OpenDaylight.

Tech Stack: Mininet, OpenDaylight, Python, Linux, SDN

Robotic Design Competition — InnovaBot

Dec 2023

Robot built on the Arduino Mega 2560 platform capable of navigating complex environments and interacting with objects, featuring sound sensitivity, color detection, maze traversal, and obstacle avoidance even in the presence of guard robots.

Tech Stack: Embedded C/C++, Arduino Mega 2560

High-Frequency Audio Power Amplifier

Dec 2023

Designed and implemented a high-frequency amplifier capable of cleanly driving an 8 Ω loudspeaker within a 20 kHz to 100 kHz frequency range, as part of a collaborative team project.

The system was built in two stages: a Class A pre-amplifier for voltage gain and a Class AB power amplifier with voltage buffers for efficient high-power delivery while maintaining linearity across the target bandwidth.

Key Results: Open-circuit gain 21, gain with 8 Ω load 16.39, bandwidth 700 kHz, input resistance 2.188 kΩ, output resistance 8.01 Ω, and total power dissipation of 2.4 W.

Head Counter System

Jul 2023

Designed and implemented an innovative "Head Counter" system capable of accurately tallying the number of individuals within a room using sensor-based and algorithmic approaches.

Tech Stack: C++, Arduino programming