Bridging the gap between AI and robotics — turning intelligent models into real-world action. Specializing in Computer Vision, AI, and Edge Computing.
Developing real-time vision systems, implementing neural networks on resource-constrained devices, and optimizing AI models for edge computing.
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 in high-performance teams, transforming innovative ideas into practical, intelligent embedded systems.
Specialized in Embedded Systems, AI, and Computer Vision
Diploma in IT & Diploma in English
Technical stack & proficiencies
Selected output & engineering solutions
Intelligent voice-driven assistant combining Agentic RAG with speech recognition. Bridges LLMs with real-world knowledge bases.
Comparative study fine-tuning large language models for a home health assistant on resource-constrained hardware.
Real-time surveillance integrating ArcFace, custom YOLOv11 for weapon detection, and a multi-agent LangGraph chatbot with ESP32 alerts.
Custom-built autonomous robot for fire detection and suppression. Complete electrical, electronic, and logic design.
IoT driver safety system with YOLO11x phone detection & MediaPipe drowsiness monitoring running at 30+ FPS, linked to ESP32.
Hybrid ensemble face system using YOLOv11 and ArcFace. Real-time 20+ FPS performance on 200k FAISS indexed embeddings.
Autonomous smart car controlled by hand gestures with real-time people tracking, combining YOLO11x-pose estimation, OpenCV, and ESP32 for human-robot interaction.
Gesture-controlled fan system using a neural network trained with Teachable Machine, recognizing five hand gestures to control power and speed.
Real-time face detection, registration, and recognition system storing embeddings in PostgreSQL for scalable identity management and fast retrieval.
Robust real-time identity verification system built on MTCNN detection and FaceNet recognition, supporting webcams through Intel RealSense D435i depth cameras.
Deep learning system using YOLOv8 to identify potential cancerous regions in medical images, assisting early detection through automated screening.
YOLOv8-based currency recognition system identifying Sri Lankan coin denominations from Rs.1 to Rs.5000, suited for vending and retail automation.
Full-stack web application for object detection powered by YOLOv8 with a Flask backend, supporting image upload, processing, and batch analysis.
Image segmentation system trained on 1,000 Roboflow-annotated images, delivering precise real-time apple detection masks for smart agriculture.