Enthusiastic AI/ML Engineer
Data-Driven Problem Solver
Passionate About Optimizing Manufacturing Processes with Machine Learning and AI
My Projects
As a passionate AI/ML developer, I’ve worked on diverse, real-world applications spanning machine learning, computer vision, robotics, and public health. My projects include everything from designing autonomous hexapod robots and applying advanced image stylization techniques, to developing real-time face mask detection systems to support global health efforts. Each project reflects my problem-solving mindset, technical depth, and drive to create impactful, intelligent solutions.
Browse below to explore how I turn ideas into practical innovation.
Building Age Recognition - AI Model
Group Project
This project presents an innovative deep learning approach to automatically predict the age of buildings using satellite images. The system follows a two-step process: first, it identifies age-related visual patterns at the image patch level; then, it aggregates these insights to generate a final prediction for entire buildings. The model was trained and evaluated on diverse datasets, with extensive testing to analyze parameter sensitivity and the performance of deep neural networks. The results demonstrate high prediction accuracy and showcase the potential of AI in automating urban analysis and architectural assessment.
Simulation of Hexapod Robot
Group Project
This project involves the end-to-end development of a hexapod walking robot using MATLAB tools. Inspired by insect locomotion, the robot features six legs and is modeled dynamically using the Simscape toolbox, with immersive 3D visualization enabled via Virtual Reality Modeling Language (VRML). A Proportional-Integral-Derivative (PID) controller is implemented to achieve precise and stable motion control, rigorously validated through simulations. The modular architecture of the dynamic model enables easy integration of advanced nonlinear control strategies, making it a versatile platform for both robotics research and biological locomotion theory testing.
Neural Style Transfer
Group Project
The motivation behind our computer vision project on image stylization is to harness the transformative power of visual aesthetics. By implementing stylization techniques, we aim to facilitate the creation of visually striking and artistic images that go beyond mere representation. Our goal is to evoke a spectrum of emotions, convey distinct styles or moods, and offer fresh perspectives on familiar content. This endeavor not only explores the intersection of technology and creativity but also opens avenues for innovative expression, allowing individuals to engage with images in novel and captivating ways.
Countour Detection
Individual Self-learning Project
This project focuses on implementing contour detection techniques in computer vision to accurately identify and extract object boundaries within images. By exploring both classical and modern algorithms, the system supports key tasks such as object recognition, image segmentation, and shape analysis. The goal is to develop a robust and adaptable framework that enhances the precision and reliability of computer vision applications across diverse domains.
Face Mask Detection
Group Project
This project implements an efficient, real-time face mask detection system using the lightweight MobileNetV2 architecture. Designed for speed and accuracy, the model analyzes facial features to determine mask presence, combining contour detection with robust classification. Its optimized performance allows deployment on edge devices and mobile platforms, making it a practical solution for public health compliance in real-world environments. The system supports global safety initiatives by enabling scalable and responsive mask enforcement technologies.
