EMBEDDED SYSTEMS ENGINEER

Specializing in Computer Vision, AI, ML & Deep Learning. Bridging the gap between hardware and intelligent systems with cutting-edge embedded solutions.

ABOUT

A committed, passionate, and talented engineer with technical expertise in Computer Vision, Artificial Intelligence, Machine Learning, and Deep Learning. Equipped with the capacity to lead and collaborate with others in a team.

My expertise lies in developing real-time vision systems, implementing neural networks on resource-constrained devices, and optimizing AI models for edge computing. I'm passionate about transforming innovative ideas into practical, intelligent embedded systems.

Professional Photo

EDUCATION

BSc (Hons) in Data Science

SLTC University

2021 - 2025

Specialized in Embedded Systems, AI, and Computer Vision

Professional Diplomas

ESOFT Metro Campus

2017 - 2018

Diploma in IT & Diploma in English

Complete School Education

Ananda College

Grade 1 - 13 | O/L & A/L

Engineering Technology, Science for Technology, Information and Communication Technology

WORK EXPERIENCE

Spera Labs

AI Engineering Intern

6 Months Internship

  • Specialized in AI-driven solutions: Speech AI, NLP, and AI Agents
  • Developed Face & Voice Recognition, 3D Face Reconstruction systems
  • Implemented AI Image Generation and Low-Quality Image Enhancement
  • Created intelligent Conversational AI chatbots with memory
  • Optimized Large-Scale Information Retrieval systems
VISIT SITE

TECHNICAL PROFICIENCY

AI, ML & Embedded Computer Vision

AI, Machine Learning & Deep Learning (Embedded Focus)
Embedded Computer Vision Systems
YOLO-Based Real-Time Object, Face & Action Detection
Image Acquisition (USB, CSI, ESP32-CAM)
Vision Model Optimization for Low-Power Devices
FPS vs Accuracy Trade-off Analysis
Camera Pipeline Optimization for Low Latency

Computer Vision

Image Acquisition (USB / CSI / ESP32-CAM)
Model Quantization (INT8 / FP16)
Memory & Compute Optimization for Vision Models
FPS vs Accuracy Trade-off Analysis
Camera Pipeline Optimization for Low Latency

Embedded Systems

Embedded C/C++ on ESP32 & Arduino
Real-Time Control Using Interrupts & Timers
Sensor, Motor & Relay Interfacing
Power Regulation, Fuse Protection & Safety Design
Embedded AI System Integration

Edge AI & Model Deployment

Edge AI & TinyML Deployment
AI Model Deployment on ESP32 & Raspberry Pi
TensorFlow Lite & TensorFlow Lite Micro
PyTorch (Model Training)
ONNX-Based Model Conversion
Model Quantization & Optimization (INT8 / FP16)

Communication Protocols

UART, I2C, SPI, PWM
Wi-Fi & TCP/IP Networking
HTTP/HTTPS & Secure MQTT (MQTTS)
Cellular Communication (2G / 3G / 4G)
LoRa / LoRaWAN
Camera & Peripheral Interfaces

Hardware & Installation

Embedded Hardware Assembly & System Integration
Panel Wiring, Cable Management & Electrical Safety
Power Distribution, Fuses & Grounding
Sensor & Camera Installation
System Testing, Commissioning & Maintenance

CERTIFICATIONS

Click to View Full Certificate

Fundamentals of Robotics & Industrial Automation

Coursera / L&T Edutech

2025

Click to View Full Certificate

Introduction to Internet of Things

Cisco Networking Academy

2025

Click to View Full Certificate

Linear Algebra for Machine Learning and Data Science

Coursera / DeepLearning.AI

2023

Click to View Full Certificate

Process Data from Dirty to Clean

Coursera / Google

2022

Click to View Full Certificate

Advanced Algorithms and Complexity

Coursera / UC San Diego

2022

Click to View Full Certificate

Capstone: Applying Project Management in the Real World

Coursera / Google

2022

Click to View Full Certificate

Google Project Management

Coursera / Google

2022

Click to View Full Certificate

Introduction to Data Science

Coursera / IBM

2022

PROJECTS

Autonomous Smart Firefighting Robot for Sri Lanka (Ongoing - 60% Complete)

Developing an autonomous firefighting robot that can move to different locations, detect fire automatically, and spray water without human intervention. All electronics and electrical system design developed through self-learning, including circuit design, power management, and system integration. Future phases will include camera-based navigation and vision systems for improved fire localization and autonomous movement.

Robotics Fire Detection Autonomous Systems Embedded Systems IoT
View Project Details

Real-Time Gender Classification Using Deep Metric Learning

A hybrid ensemble system combining YOLOv11 for face detection, ArcFace (ResNet-50) for facial embedding extraction, and dual classification using FAISS k-NN (91.38% accuracy, 97.63% AUC) and MLP neural network (90.05% accuracy). Achieved real-time performance at 20+ FPS on GPU hardware with 200,046 face embeddings, demonstrating effective deep metric learning with vector database indexing.

Deep Learning YOLOv11 ArcFace FAISS Real-Time Processing
View on GitHub

CRIMEGUARD: AI Security System for Crime Detection (Final Year Project)

Intelligent surveillance system integrating InsightFace ArcFace for face recognition (95% accuracy), custom YOLOv11 for weapon detection (88% precision, <480ms latency), and multi-agent chatbot (LangGraph + GPT-4o-mini + Gemini) supporting Sinhala, Tamil, and English. Features ESP32-based physical alerts and Twilio communication for real-time emergency response in Sri Lankan security applications.

YOLOv11 ArcFace LangGraph ESP32 Real-Time AI
View on GitHub

Driver Safety Guard 2.0: Real-Time IoT Safety Monitoring

AI-powered driver safety system combining YOLO11x for phone detection (94.2% accuracy) and MediaPipe for drowsiness monitoring, achieving 30+ FPS real-time processing. Features ESP32-based alerts (RGB LEDs, audio warnings), Blynk mobile app integration, and GPU optimization. Collaborative project demonstrating practical AI application for proactive road safety in Sri Lankan context.

YOLO11x MediaPipe ESP32 IoT Computer Vision
View on GitHub

Vision Guided Smart Car with AI Hand Gesture Control

Intelligent autonomous car system responding to hand gestures with active people tracking using advanced computer vision. Combines YOLO11x-pose estimation for gesture recognition, OpenCV for real-time processing, and ESP32 microcontroller for seamless human-robot interaction and vehicle control.

YOLO11x-pose OpenCV ESP32 Gesture Recognition Autonomous Systems
View on GitHub

Smart Fan Control Using Image Processing

AI-powered gesture recognition system using neural networks trained with Google's Teachable Machine to control fan operations. Recognizes five distinct hand gestures for turning the fan on/off, increasing/decreasing speed, and setting specific speed levels. Demonstrates practical application of machine learning for smart home automation and gesture-based control systems.

Machine Learning Gesture Recognition Teachable Machine Image Processing Smart Home
View on GitHub

Face Recognition System Using Aiven PostgreSQL

Powerful and efficient face recognition system built with OpenCV, imgbeddings, and PostgreSQL for real-time face detection, registration, and recognition. Features database storage for embeddings, enabling scalable identity management and fast retrieval for real-time applications.

OpenCV PostgreSQL Face Recognition imgbeddings Database
View on GitHub

Face Recognition System Using MTCNN and FaceNet

Comprehensive face recognition system implementing cutting-edge MTCNN face detection and FaceNet recognition algorithms for robust, real-time identity verification. Supports multiple hardware configurations from standard webcams to Intel RealSense D435i depth cameras, suitable for both basic and advanced security applications.

MTCNN FaceNet Real-Time Intel RealSense Security
View on GitHub

Advancing AI in Healthcare: Breast Cancer Detection with YOLOv8

Deep learning-powered breast cancer detection system using YOLOv8 object detection model to identify potential cancerous regions in medical images. Leverages computer vision to assist in early detection, significantly improving survival rates through AI-assisted medical image analysis and automated screening.

YOLOv8 Healthcare AI Medical Imaging Cancer Detection Computer Vision
View on GitHub

Sri Lankan Coin Detection System

Intelligent coin detection system using YOLOv8 to accurately identify and categorize Sri Lankan currency coins across all denominations (Rs.1 to Rs.5000). Demonstrates practical application of computer vision for automated currency recognition, suitable for vending machines, cash handling systems, and retail automation.

YOLOv8 Object Detection Currency Recognition Computer Vision Automation
View on GitHub

Detective Lens

Web-based object detection application powered by YOLOv8 with Flask backend for image upload and processing. Features organized directory management for results, enabling efficient batch processing and analysis. Demonstrates full-stack integration of deep learning models for accessible computer vision applications.

YOLOv8 Flask Web Application Object Detection Full-Stack
View on GitHub

Apple Detection System Using YOLOv8x-seg

Advanced image segmentation system using YOLOv8x-seg trained on 1,000 annotated images via Roboflow over 50 epochs. Achieves accurate real-time apple detection and segmentation with precise object masks. Ideal for smart agriculture applications, quality control, and automated fruit sorting systems.

YOLOv8x-seg Segmentation Roboflow Agriculture AI Computer Vision
View on GitHub

GET IN TOUCH

CONTACT INFO

I'm always interested in discussing new projects, innovative ideas, or opportunities to be part of cutting-edge embedded AI solutions.

×