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"
};
-
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.