Engineering Explainable AI Systems Where Accuracy Alone Is Not Enough.
I am a Master’s student in Artificial Intelligence, specializing in Explainable AI (XAI) and Quantum Machine Learning. My core passion lies in breaking down the "black box" of complex deep learning models—specifically for critical healthcare systems and diagnostic transparency. Parallelly, I am an experienced Flutter Developer with a production-first mindset for building scalable mobile systems.
- Beyond the Black Box (XAI): Utilizing feature attribution methods (SHAP, LIME, Integrated Gradients, Grad-CAM) to make medical diagnostics interpretable for clinicians.
- Quantum Machine Learning: Developing hybrid quantum-classical neural networks to optimize high-stakes predictive systems.
- Trustworthy Decision Frameworks: Combining advanced deep learning (GNNs, Transformers) with interpretable algorithms like Fuzzy Logic.
| Domain | Tools & Frameworks |
|---|---|
| Explainable AI (XAI) | SHAP • LIME • Integrated Gradients • Grad-CAM • Fuzzy Logic |
| Deep & Machine Learning | PyTorch • TensorFlow • Keras • Scikit-Learn • GNNs • LSTMs |
| Quantum Computing | PennyLane • Hybrid QCNNs |
| Data Science & Analytics | Python • Pandas • NumPy • Matplotlib • Seaborn |
| Mobile Engineering | Flutter • Dart • Firebase • Supabase • Bloc/Provider • REST APIs |
Research Paper
A Multi-Method Explainability Study of a Hybrid Quantum-Classical Neural Network for Liver Disease Detection
This research investigates the integration of Hybrid Quantum-Classical Neural Networks with Explainable AI (XAI) techniques for trustworthy clinical decision support.
The study explores:
- Hybrid quantum-classical learning architectures using PennyLane and TensorFlow
- Multi-method explainability analysis using SHAP and Integrated Gradients
- Feature-level interpretation for healthcare decision transparency
- Trustworthy AI evaluation for high-stakes medical prediction systems
🔬 Code Repository:
https://github.com/fatimasood/XAI-Hybrid-Quantum-Liver-Disease-Detection
Research Area: Explainable AI • Quantum Machine Learning • Healthcare AI • Trustworthy AI
- XAI Hybrid Quantum Liver Disease Detection: Research-driven hybrid quantum-classical neural network for liver disease prediction, integrating PennyLane, TensorFlow, SHAP, and Integrated Gradients for interpretable healthcare AI. This work is associated with an academic preprint on trustworthy quantum machine learning for clinical diagnostics.
- CNN-Based Early Autism Detection: Utilizing facial image analysis with advanced CNN architectures (Xception, VGG16) integrated into a reproducible ML pipeline for early ASD biomarkers.
- XAI Diabetes Prediction Engine: A Stacking Ensemble classifier mapped with LIME to provide transparent, feature-level insights into patient data.
- NeuroVerify Engine: A multimodal fake news detection pipeline combining LSTMs, Graph Neural Networks (GNNs), and Transformers with a novel rule-based Fuzzy Logic system for interpretable trust scoring.
- High-Resolution Face Restoration: A deep learning pipeline featuring U-Net Autoencoders, GANs, and Hybrid Context Encoders for restoring damaged portraits.
- Smart Helmet Mobile App: A real-time IoT-integrated Flutter application engineered for instant accident detection and emergency aid deployment.
- InfoKlub App: A secure personal data ecosystem featuring drag-and-drop file uploads, AI-powered CV generation, and end-to-end data encryption.
- 🔭 Currently refining: Quantum AI frameworks and human-interpretable validation models.
- 💡 Ask me about: Why your model acts like a black box and how we can add transparency layers to it.
- 🤝 Open to: Academic research collaboration, XAI integrations, and architectural consulting.
"I believe AI systems used in healthcare, security, and public decision-making must prioritize interpretability, accountability, and human trust—not just predictive accuracy."