Luxshan Thavarasa
Software Engineer @ H2O.ai | Research Enthusiast | CSE @ University of Moratuwa
Location: Visuvamadu East, Visuvamadu, Mullaitivu, Northern Province, Sri Lanka
Email: luxshan2327 [AT] gmail.com
About
Final-year Computer Science and Engineering student at the University of Moratuwa with a passion for Full Stack Development and Artificial Intelligence. My research focuses on Speech Emotion Recognition (SER) for low-resource languages, aiming to bring AI advancements to diverse linguistic communities.
I enjoy working in collaborative environments and thrive on creating innovative solutions that make complex data accessible and meaningful to users. I'm eager to connect with like-minded individuals and work on projects that push the boundaries of technology.
Research Focus: Speech Emotion Recognition · Natural Language Processing · Agentic AI Systems · Full-Stack Development
Languages: Tamil (Native), English (Professional Working)
Current Role: Software Engineer at H2O.ai, developing frontend and backend solutions for H2OGPTe and Agentic Applications.
Experience
Professional
Software Engineer, H2O.ai (Colombo, Sri Lanka)
H2OGPTe: Developing the frontend with a strong focus on UI/UX for agent interactions, while contributing to backend improvements to ensure seamless functionality.
Intern Engineering - Software, H2O.ai (Colombo, Sri Lanka)
- H2OGPTe: Developed frontend with focus on UI/UX for agent interactions while contributing to backend enhancements for seamless functionality
- Agentic Application: Worked on both frontend and backend development using React with Fluent UI and H2O UI tools for frontend, and FastAPI with PostgreSQL for backend
- ChurnApp: Created application to predict customer churn for H2O.ai. Trained model with H2O Driverless AI and deployed in MLOps environment. Integrated Salesforce and helpdesk data with React frontend and FastAPI backend
- Olympic App: Developed platform for conducting hackathons, including participant and administrative applications using H2O Wave and PostgreSQL
Full-stack Developer, aaivu (Colombo, Sri Lanka)
Rtuthaya Website: Contributed to the development of Dr. Uthayasankar Thayasivam's website. Managed content creators and guided new juniors in enhancing the application. Key contributions included project page and conference page development using PHP, MySQL, HTML, CSS, JavaScript, and Bootstrap.
Education
BSc Eng Hons, Computer Science & Engineering
University of Moratuwa, Sri Lanka
- Specialization: Computer Science and Engineering
- Key Skills Developed: Project Documentation, Computer Science, MySQL, Flask, NumPy, Software Development, Problem Solving, Software Architecture, Python, Data Structures, Java, Web Services, Mathematics, Solution Architecture, Software Quality, Technical Documentation, Analytical Skills, Leadership, Test Cases, API Development, SQL, RDBMS, Pandas, Communication, Test Planning, Scripting, Architecture Development, Generative AI, C++, Unit Testing
- Final Year Project: Multilingual Universal Speech Emotion Recognition Model
Secondary Education
Mu/Visuvamadu Maha Vidyalayam, Mullaitivu, Sri Lanka
G.C.E. (A/L) Examination 2019: 3A grades | Physical Science stream | Island Rank: 209 | Z-Score: 2.4704
G.C.E. (O/L) Examination 2016: 8A, 1B | Including Mathematics, Science, English, and Business & Accounts
Primary Education
Kn/Tharmapuram G.T.M.S, Sri Lanka
Publications
This study introduces a novel multilingual model designed to effectively address the challenges of detecting abusive content in low-resource, code-mixed languages. Our model achieved a macro F1 score of 0.7864 on Tamil dataset and 0.7058 on Malayalam dataset using transfer learning techniques and multi-head attention mechanisms.
This paper introduces EmoTa, the first emotional speech dataset in Tamil, designed to reflect the linguistic diversity of Sri Lankan Tamil speakers. EmoTa comprises 936 utterances from 22 native Tamil speakers across five primary emotions with substantial Fleiss' Kappa agreement score of 0.74 and F1-scores of 0.91 and 0.90.
Key Projects
DravidaKavacham – Abusive Content Detection for Tamil & Malayalam
Developed an open-source tool for detecting abusive Tamil and Malayalam content targeting women. Built as part of DravidianLangTech @ NAACL 2025, leveraging NLP and machine learning for accurate text classification.
Technologies: Python, Machine Learning, Transfer Learning, Attention Mechanisms
QuickChat - Chrome Extension Help Chatbot
Developed Chrome extension integrating help chatbot into any website's bottom right corner. Backend processes involve web scraping, content embedding with OpenAI embeddings and Chroma DB, with real-time responses using GPT-3.5.
Technologies: JavaScript, HTML/CSS, Python, FastAPI, Web Scraping, OpenAI Embeddings, Chroma DB, Vector Databases, RAG
Driver's License Test Application
Developed user-friendly web portal and mobile application for driver's license exam preparation with comprehensive study materials, practice tests, and performance feedback.
Technologies: HTML, CSS, JavaScript, Bootstrap, React JS, Express JS, MongoDB, React Native, Git, CI/CD, Docker, AWS EC2, Redux.js
Supply Chain Management System
Developed web-based SCMS for online shopping with team of 5 members. Designed RDBMS for efficient data management and implemented triggers and functions for enhanced database performance.
Technologies: HTML, CSS, JavaScript, Bootstrap, PHP, MySQL
Email Client
Developed console-based Email Client application using Java programming, design patterns, and OOP concepts. Features include contact management, email sending, automated birthday wishes, and email history.
Technologies: Java, Object-Oriented Programming, Design Patterns
Technical Skills
Programming Languages: Python, JavaScript, TypeScript, Java, C++, PHP, HTML/CSS
Frameworks & Libraries: React, FastAPI, Express.js, Flask, Node.js, React Native, Bootstrap, jQuery, PyTorch, NumPy, Pandas
Databases: MySQL, PostgreSQL, MongoDB, Chroma DB, RDBMS
Cloud & DevOps: AWS (EC2, S3, RDS), Docker, GitHub Actions, CI/CD, Nginx, Reverse Proxy
AI/ML Technologies: Machine Learning, Deep Learning, PyTorch, Transfer Learning, NLP, Large Language Models, Generative AI, CNN, Attention Mechanisms
Development Tools: Git, GitHub, REST APIs, OAuth 2.0, Unit Testing, Jest, Version Control, Web Scraping
Specializations: Full-Stack Development, Speech Emotion Recognition, Natural Language Processing, Agentic Systems, UI/UX Design
Honors & Awards
- All Island Mathematics Competition 2018: Senior 2, 2nd Runner-Up (Northern Province Team) - Ministry of Education, Sri Lanka
- Academic Excellence: Island Rank 209 in G.C.E Advanced Level with Z-Score 2.4704
- Research Publications: Accepted papers in COLING 2025 and NAACL 2025 workshops
- Professional Recognition: Promoted from Intern to Software Engineer at H2O.ai
Licenses & Certifications
Fundamentals of Deep Learning - NVIDIA (March 2025)
Credential ID: 1tO0Ys3ITkGJkXM3sgBKrQ
Skills: Deep Learning, Machine Learning, PyTorch, Neural Networks
Large Language Models (LLMs) - Level 1 - H2O.ai (March 2024)
Skills: Large Language Models, Generative AI, Natural Language Processing
Machine Learning A-Z - Udemy (March 2024)
Credential ID: 60b2314b-e9de-4b14-af89-b301ec16a5ed
Skills: CNN, Deep Learning, Supervised/Unsupervised Learning, PyTorch, NLP
Meta Frontend Development Specialization - Coursera (2023)
Comprehensive Program Including: Advanced React, React Basics, HTML and CSS in depth, Programming with JavaScript, UX Design Foundations, Version Control
Volunteering
Reviewer - DravidianLangTech Workshop (NAACL 2025)
Contributed as volunteer paper reviewer for the Fifth Workshop on Speech, Vision, and Language Technologies for Dravidian Languages. Evaluated research submissions specializing in hate speech detection across text and speech data.
Contact Me
Get in Touch
Email: luxshan2327@gmail.com
Mobile: +94-77-3038402
Address: Visuvamadu East, Visuvamadu, Mullaitivu, Northern Province, Sri Lanka
LinkedIn: linkedin.com/in/lux-thavarasa
GitHub: github.com/luxshan2000
Languages: Tamil (Native), English (Professional Working)
"Let's shape the future together! I'm eager to connect with like-minded individuals and work on projects that push the boundaries of technology."