CV
Amin Jalali
Summary
Postdoctoral Fellow at Queen's University developing AI pipelines for epilepsy detection and presurgical decision-making using advanced neuroimaging and foundational models
Education
- Ph.D in Electronic & Electrical Engineering2021-08-01Kyungpook National UniversityGPA: Best Thesis AwardCourses: Low-shot, long-tailed learning for medical imaging, Time-series prediction, Deep learning for medical applications
- M.S. in EngineeringUniversity
- B.S. in EngineeringUniversity
Work Experience
- Postdoctoral Fellow2023-04-01 -Queen's UniversityCentre for Neuroscience Studies & Ingenuity Labs Research Institute
- Developing AI pipelines for epilepsy detection and presurgical decision-making using arterial spin labeling MRI
- Building foundational models for neurological disease prediction
- Working on hierarchical time-series representation learning and multi-domain EEG modeling for cognitive-load classification
- Postdoctoral Research Fellow2021-09-01 - 2023-03-01KNU-LG Convergence Research CenterSouth Korea
- Built Transformer-based ICU outcome predictors
- Developed AI clinical-decision support systems
- Advanced multimodal medical AI research
Skills
Deep Learning & AI
- CNNs
- Transformers
- Self-supervised Learning
- Contrastive Learning
- Foundation Models
Medical AI
- MRI Analysis
- EEG Processing
- Time-series Forecasting
- Anomaly Detection
- Clinical Decision Support
Machine Learning
- Long-tailed Learning
- Low-shot Learning
- Meta-learning
- Imbalance Regularization
Computer Vision
- Medical Imaging
- Hippocampal Sclerosis
- Arterial Spin Labeling
- Image Analysis
Programming
- Python
- PyTorch
- TensorFlow
- MATLAB
- R
Publications
- Adaptive Metadata-Guided Supervised Contrastive Learning for Epilepsy Detection2025IEEE Journal of Biomedical and Health InformaticsNovel approach for epilepsy detection using metadata-guided contrastive learning methods with advanced MRI analysis.
- Dynamically Adaptive Deformable Feature Fusion for Medical Imaging2025Engineering Applications of Artificial IntelligenceAdvanced feature fusion techniques for improved medical imaging analysis with deformable architectures.
- Learnable Feature Alignment with Attention-Based Data Augmentation for Medical Time Series2024Applied Soft ComputingAttention-based data augmentation methods for medical time series analysis with learnable feature alignment.
- Low-Shot Imbalanced Data Regularizations for Medical Imaging2021PhD Thesis - Kyungpook National UniversityComprehensive study on regularization techniques for low-shot and imbalanced medical imaging datasets. Awarded Best Thesis.
Presentations
- AI Pipelines for Epilepsy Detection: From MRI to Clinical Decision Support2025Queen's University, Centre for Neuroscience StudiesKingston, Ontario, CanadaInvited talk on advances in AI pipelines for epilepsy detection using arterial spin labeling MRI
- Hierarchical Time-Series Representation Learning for Medical AI2024International Conference on Medical AISeoul, South KoreaConference presentation on hierarchical time-series methods for EEG and ICU prediction
- Adaptive Contrastive Learning for Neurological Disease Prediction2024IEEE International Conference on Biomedical and Health InformaticsChicago, IL, USAResearch presentation on metadata-guided contrastive learning approaches
- Low-Shot Learning for Medical Imaging: Theory and Practice2022KNU-LG Research SymposiumDaegu, South KoreaKeynote presentation on low-shot learning techniques for medical imaging applications
Teaching
- Deep Learning for Medical Applications2023Queen's University, Department of Electrical and Computer EngineeringRole: Graduate Course InstructorAdvanced graduate course covering deep learning applications in medical imaging and biomedical signal processing
- AI Workshop: Medical Image Analysis2022KNU-LG Convergence Research CenterRole: Workshop InstructorIntensive workshop on AI techniques for medical image analysis with hands-on implementation sessions
Portfolio
- AI Pipeline for Epilepsy Detection2025PortfolioAdvanced AI system for epilepsy detection using arterial spin labeling MRI and foundational models
- Hierarchical Time-Series Representation Learning2024PortfolioMulti-domain EEG modeling for cognitive-load classification and ICU outcome prediction
Interests
- Deep Learning & Foundation ModelsCNNs, Transformers, Self-supervised Learning
- Medical & Biomedical AIMRI, EEG, Time-series Forecasting, Anomaly Detection
- Long-Tailed & Low-Shot LearningImbalance Regularization, Meta-loss Methods
- Computer Vision & Medical ImagingHippocampal Sclerosis, Arterial Spin Labeling