I am an Artificial Intelligence postgraduate student at Macquarie University and a Computer Science and Engineering graduate from BRAC University.

• 🔭 I am building AI projects in NLP, computer vision, graph machine learning, and intelligent systems.

• 🌱 I am currently exploring deep learning, explainable AI, LLM applications, and robot vision.

• ⚡ I enjoy turning research ideas into working, well-documented machine learning projects.

• 🎓 I previously worked as a part-time Lecturer and Student Tutor at BRAC University.

const Joya = {
 pronouns: "she/her",
 education: [
 "MIT in Artificial Intelligence - Macquarie University",
 "BEng in Computer Science and Engineering - BRAC University"
 ],
 code: ["Python", "C", "C++", "Java", "SQL"],
 ai: [
 "Deep Learning",
 "Computer Vision",
 "Natural Language Processing",
 "Graph Neural Networks",
 "Explainable AI",
 "LLM Applications"
 ],
 tools: [
 "PyTorch",
 "HuggingFace Transformers",
 "PyTorch Geometric",
 "spaCy",
 "SHAP",
 "ROS 2",
 "Scikit-learn"
 ],
 interests: ["AI research", "teaching", "robot vision", "real-world ML systems"],
 currentFocus: "Building practical AI projects with clear explanations and reproducible workflows"
};

Python PyTorch HuggingFace Scikit-learn NumPy Pandas OpenCV Jupyter C C++ Java MySQL LaTeX Linux

• Fine-Grained Beverage Can Classification for ROS2-Based Robotic Perception
  Computer vision and robotics project using MobileNetV3-Large, YOLOv8 crop detection, PyTorch, OpenCV, and ROS2 for real-time beverage can classification and robot target-search behaviour. Achieved 97.20% accuracy on 20 classes and 100% accuracy on the 3-class target-search subset.

• Multi-Generator Deepfake Document Detection and Tamper Localization
  End-to-end document forensics project for detecting AI-generated/deepfake documents and localizing tampered regions using computer vision, PyTorch, and image segmentation techniques.

• LLM-Powered Personal Travel Assistant
  End-to-end LLM application for automated travel planning using prompt engineering, dialogue management, and personalized text generation.

• OSHC Insurance Claim JDM Rules
  Business-rule decision model for Overseas Student Health Cover insurance claims using GoRules/JDM decision tables and a Python reference simulator.

• Feedback Analysis for Products Using NLP
  Aspect-based sentiment analysis on Amazon reviews using BERT and SHAP explainability for transparent product feedback insights.

GitHub Streak

Top Langs

trophy