NIMSHA FERNANDO
I'm Nimsha, a software engineering undergraduate from Sri Lanka with a strong focus on computer science research and intelligent systems. My work spans machine learning, applied AI, and full-stack development — from ensemble CNNs and NLP pipelines to cloud-native architectures. I'm driven by rigorous research that bridges theory and real-world impact.
A small selection of recent projects
MintWatches
Built a full-stack Flutter (Dart) e-commerce app integrated with a RESTful Laravel / MySQL backend. Developed a custom admin panel with role-based access control for product and order management.
Petfinity
Developed a Laravel MVC pet-care platform featuring owner profiles, bookings, and an analytics dashboard. Managed the project end-to-end via Agile/Scrum ceremonies tracked on Jira.
Morse Code Translator
Built a bidirectional Morse translator using a binary tree data structure for O(log n) character lookup. Added a message-reversal obfuscation layer, designed under clean OOP principles.
Inventory Management System
Built an inventory management system with a external transaction log using C# and MySQL.
Responsive Turtle Conservation Website
Built a responsive multi-page conservation site with interactive UI components using no external frameworks. Implemented a vanilla-JavaScript ticket booking and seat-management system.
Portfolio Website
Built an interactive Next.js portfolio featuring Three.js 3D WebGL animations and Framer Motion page transitions. Deployed on Vercel with an automated CI/CD pipeline for continuous integration and delivery.
PhytoSense
Built an ensemble CNN (ResNet-50, MobileNetV2, EfficientNet-B0) reaching 82.39% accuracy across 80 medicinal plant species, plus a QSAR XGBoost model on ∼19,800 ChEMBL EGFR compounds (R² = 0.90). Integrated RDKit Morgan fingerprints for phytochemical retrieval and deployed via Flask with GPT-4-based explainability descriptions.
SYOS
Built a multithreaded Java POS and inventory system applying concurrency control and Gang-of-Four patterns under SOLID principles. Engineered thread-safe synchronization and validated via a JUnit 5 / Mockito suite with SonarQube static analysis.
MDS
Fine-tuned YOLOv8s on MosquitoFusion (1,204 images) for 3-class real-time detection of mosquitoes, swarms, and breeding sites. Deployed via a Gradio interface generating automated risk-level reports from detected objects.
CardioRisk AI
Built an end-to-end ML pipeline on the Framingham dataset (4,240 patients) training LR, RF, and XGBoost with SMOTE for class imbalance (ROC-AUC ∼0.79). Deployed an interactive Gradio model-comparison UI for real-time 10-year coronary heart disease risk prediction.
Sinha-Sign
Fine-tuned InceptionV3 (∼93% val accuracy) integrated with an OpenCV DNN hand detector for real-time gesture recognition. Built a Gradio UI outputting Sinhala script, romanisation, and gTTS-generated spoken audio.
Skin-Oil
Built a K-Means pseudo-labelling pipeline using HSV / YCbCr features extracted via CLAHE for unlabelled skin images. Fine-tuned MobileNet to classify four oiliness levels, deployed with in-browser webcam capture in Gradio.
APIIT Intern Management System
Built a multi-role Laravel platform (Student, Company, Industry Liaison, Admin) for end-to-end internship lifecycle tracking. Implemented an analytics dashboard, role-based access control, and CSV/Excel import/export functionality.
SynapseSafe
Built a hybrid desktop + web platform performing real-time, on-device screen monitoring with a custom CNN (94%+ accuracy, < 5% CPU) and context-aware educational intervention. Developed a Node.js dashboard with CRUD admin panel, session-based auth, and a GPT-4 / Groq-powered chatbot supporting Sinhala, Tamil, and English.
CloudRetail
Designed a fault-tolerant AWS microservices architecture (RDS Multi-AZ, auto-scaling ECS, CloudFront CDN) with serverless Lambda / EventBridge pipelines. Achieved 100% pass rate across 134 automated unit, integration, and load tests with academic documentation on cost and fault-tolerance.
CodeClarity
Built a VS Code extension for semantic readability analysis – camelCase/snake_case identifier tokenisation, curated SE-vocabulary validation, and scope-aware comment alignment with a GPT-assisted synonym-detection layer. Aggregated naming, comment, and lexical-consistency metrics into a weighted readability score, validated via Spearman correlation (ρ up to 0.82) and a 30-student user study (mean rating 4.21/5).
A showcase of my technical expertise
Frontend Development
HTML
CSS
JavaScript
TypeScript
React
Next.js
Framer Motion
Three.js
Bootstrap
Tailwind CSS
Flutter (Dart)Backend Development
Python
PHP
Java
C#
Laravel
Node.js
MySQLMachine Learning & AI
TensorFlow
PyTorch
GPT-4 / LLMsScikit-learnTools Used
AWS
Docker
VS Code
Visual Studio
GitHub
GitBash
IntelliJ IDEA
PHPStorm
Android Studio
SSMS
MySQL
JIRA
Figma
Notion
ConfluenceMy work experience
Research Assistant
APIIT Lanka (Pvt) Ltd
Colombo, Sri-Lanka
- Conducted academic research writing and systematic literature reviews using LaTeX and PRISMA methodology.
- Prepared Software Requirements Specification (SRS) documents and supported requirements elicitation for academic projects.
- Participated in research meetings, contributing to methodology discussions and academic publishing workflow.
Business Analyst (Intern)
Rootcode (Pvt) Ltd
Colombo, Sri-Lanka
- Analyzed business workflows and gathered requirements through stakeholder discussions to identify process improvement opportunities.
- Designed wireframes and AI-based prototypes, preparing comprehensive documentation including user stories and requirement specifications.
- Participated in Agile ceremonies and collaborated with QA and engineering teams to support solution implementation and testing.
Client Project Assistant
Xyicon LK (Pvt) Ltd
Colombo, Sri-Lanka
- Wrote user stories and documented feature requests in ClickUp while tracking project progress using JIRA.
- Provided customer support via live chat and HubSpot, ensuring timely issue resolutions and bug reporting.
- Participated in Azure DevOps standups and conducted product training for newly joined team members.
Client Project Assistant (Intern)
Xyicon LK (Pvt) Ltd
Colombo, Sri-Lanka
- Managed documentation to ensure accuracy and smooth operations.
- Resolved asset data issues, coordinated with the engineering team, and ensured KPIs were met.
- Designed smart kit solutions (e.g., Smart Pump Room and Electric Meter Kits) and collaborated with stakeholders to align operations with business goals.
My Education Journey
Colombo South International College (CSI College)
Completed Pearson Edexcel International GCSE & Advanced Level
Australian College of Business and Technology (ACBT)
University Diploma in Science, Computing / IT
Asian Pacific Institute of Technology (APIIT)
BEng (Hons) Software Engineering