cv

ML and R&D Engineering Intern (Remote)

Braingaze, Spain

• Developed an end-to-end ML pipeline for neurodevelopmental diagnosis using ocular biomarkers, processing raw time-series gaze data
• Designed a Deep Learning-based framework integrating Explainable AI (XAI) for clinical interpretability
• Co-authored a pending research paper on 'Clinical validation of ocular markers for differential diagnosis of ADHD and Autism'

Bachelor of Science in Engineering

University of Moratuwa, Sri Lanka

Electronic and Telecommunication Engineering

• Computer vision
• Deep Learning
• Signal processing
• Machine Learning

High school

Richmond College Galle, Sri Lanka

• Physical science

First Runners-up | Sri Lanka Circuits Challenge- 2024

IEEE Sri Lanka ChallengeSphere

Smart Glasses for Visually Impaired project, recognized in the open-ended electronic product design

First runners up | AURORA- 2024 (AI IDEATHON)

University Competition

Implementation of a Criminal Face Sketch Generator using Generative AI project Selected from University Competition Out of 47 teams

Selected for Top 30 entries | IEEE IES Generative AI Challenge- 2024

IEEE IES

Implementation of a Criminal Face Sketch Generator using Generative AI project Selected Out of 150 International teams

Finalist | IEEE ComSoc- ComFix 2024

IEEE ComSoc

Working on a solution with addressing Bandwidth Constraints in Underwater Acoustic Communication

Mahapola Higher Education (Merit) Scholarship

Government of Sri Lanka

For outstanding performance in GCE A/L Examination All Island Rank 204

Estimation of ECM Signatures in Fiber Probe Images

End-to-end ML pipeline for in-vivo extracellular matrix (ECM) quantification from Lightcore fiber-probe images.

• Addresses low resolution, noise, and optical distortions by mapping low-magnification Lightcore images to high-fidelity tabletop microscope images
• Integrates multi-magnification microscopy data (10×/20×/25×), shared-embedding learning, and uncertainty-aware scoring for accurate clinical tissue assessment
• Supervisors: Dr. Chamira Edussooriya, Dr. Ranga Rodrigo (University of Moratuwa, Sri Lanka), Dr. Dushan Wadduwage (Old Dominion University, USA), Dr. Einstein Gnanatheepam (Tufts University, USA)

Stable Diffusion Based Criminal Face Generation for Forensic Identification

Development of a Criminal Face Sketch Generator using Generative AI to generate accurate facial sketches and variations for streamlining investigations and forensics.

• Used LLMs, PyTorch, Stable Diffusion models, Hugging Face Transformers, ONNX as development tools
• Implemented approaches including Prompt Generator LLM, Stable Diffusion Model, Image-to-Image Refiner, Keyword Tracker, and Voice-based Input
• Won multiple competitions including AURORA-2024 (1st runners up) and IEEE IES Generative AI Challenge (Top 30)

Cell Anomaly Detection using Autoencoders

Implemented an autoencoder-based method in TensorFlow to reconstruct normal images and classify anomalies based on reconstruction error and Kernel Density Estimation (KDE).

• Applied insights from the paper 'Robust Anomaly Detection in Images using Adversarial Autoencoders' by Laura Beggel
• Utilized a VGG16-based model architecture for encoding and decoding images, modified to include KDE for anomaly likelihood estimation
• Enables robust anomaly detection in cell images, potentially improving medical diagnostics

Vision-Based Robot System for SLRC

Development of a machine vision system to identify 3D objects and colors using TensorFlow and OpenCV for Sri Lanka Robotics Competition.

• Fully deployed on Raspberry Pi 4 Model B using TensorflowLite
• This robot system able to secure finalist position in the competition
• Integrated computer vision for real-time object detection and classification

GenericHTTP Arduino Library

Arduino based library that simplifies sending HTTP requests using the ESP8266-01 module (ESP-01) with an Arduino based board.

• Simplified HTTP communication for Arduino projects
• Enhanced IoT connectivity for embedded systems
• Open-source contribution for Arduino community

Handwritten Digit Classification

Developed a Handwritten Digit Classification model using TensorFlow's Keras API, achieving high accuracy in classifying handwritten digits from the MNIST dataset.

• Built a neural network to classify handwritten digits from the MNIST dataset
• Preprocessing image data and encoding categorical labels for model training
• Evaluated model performance and visualized results for improved understanding
• Provided hands-on experience in image classification and model evaluation using TensorFlow

Five band Audio Equalizer

Developed a Five-Band Audio Equalizer using operational amplifier based active filters, enabling detailed audio customization across five frequency bands.

• Designed and built a fully functional audio equalizer using op-amp circuits
• Implemented low-pass, high-pass, and band-pass filters for audio signal manipulation
• Provides adjustable frequency bands for precise audio customization, including Low, Low Mid, Mid, High Mid, and High
• Enhances hands-on electronics skills with practical application of analog circuit design

Sparkle Robot

Developed a versatile robotic platform powered by Arduino Mega 2560, showcasing advanced functionalities and innovative sensor integrations.

• Implemented a range of functionalities using Arduino programming
• Integrated various sensors for enhanced robot capabilities
• Demonstrated problem-solving skills in robot design and development
• Provided hands-on experience in robotics and microcontroller programming

Automatic Retractable Clothesline System

Developed an Automatic RRetractable Clothesline System (ARCS) incorporating Arduino technology, automated rain detection, and remote monitoring through Thingspeak.

• Combines Arduino technology with automation for smart clothes drying
• Features an automated rain detection system for protecting garments
• Utilizes Thingspeak integration for real-time remote monitoring and control
• Offers an innovative solution for efficient and intelligent clothes drying

Vice President - SLRC branch

Electronic Club– University of Moratuwa

SLRC 2025/26

• Leading strategic planning for school workshops and robotics outreach
• Coordinating with university committees and external partners for event execution

Event Coordinator - SLRC Branch

Electronic Club– University of Moratuwa

SLRC 2024/25

Co-Chair | Sri Lankan Robotics Challenge 2024

Electronic Club– University of Moratuwa

As Co-Chair of the Sri Lankan Robotics Challenge 2024, I had the honor of leading the 10th edition of this prestigious national event

• Event Leadership: Organized and managed school workshops, prepared the competition arena, and ensured smooth execution of the final events
• Team Collaboration: Led a diverse team, fostering strong collaboration to execute the event's multifaceted requirements
• Innovation: Focused on promoting creativity and technical innovation among participants, inspiring them to push the boundaries of robotics and automation
• Skills Development: Enhanced my leadership, teamwork, and problem-solving skills by addressing the challenges of managing a large-scale event

SLRC Branch - Committee Member

Electronic Club– University of Moratuwa

SLRC 2023/24

Volunteer

Sasnaka Sansada Foundation