M.Sc. Artificial Intelligence & Robotics | Aspiring Ph.D. Candidate
Bridging Neurophysiology and Deep Learning to advance Non-Invasive Neural Interfaces.
My research is dedicated to Neuro-AI and Brain-Computer Interfaces (BCI). I focus on developing explainable deep learning architectures to decode complex human cognitive states from non-invasive bio-signals.
Specifically, my recent work involves mapping high-dimensional spatio-temporal EEG signals
Key Areas: Brain-Computer Interfaces (BCI) • Explainable AI (XAI) • Spatio-Temporal Transformers • Medical Image Segmentation • Autonomous Robotics
Islamic Azad University | GPA: 17.00/20.00
- Thesis: Imagined Speech Decoding from EEG Signals using Deep Learning (Grade: 20/20)
- Methodology: Proposed a novel Spectro-Temporal Transformer to capture long-range dependencies in EEG. Applied SHapley Additive exPlanations (SHAP) to validate neurophysiological fidelity and mitigate shortcut learning.
- Stack:
PyTorch,MNE-Python,SciPy,Kara One Dataset
Semnan University | GPA: 17.00/20.00
- Final Project: Design and Implementation of an Autonomous Line Follower Robot (Grade: 20/20)
- Outcome: Published as a comprehensive technical book.
| Project | Description | Tech Stack |
|---|---|---|
| Schizophrenia Detection | Developed a diagnostic pipeline mapping 1D EEG to 2D time-frequency scalograms, processed via EfficientNet. Achieved 97% accuracy in clinical validation. |
MNE-Python, PyTorch, CWT |
| Brain Tumor Segmentation | Designed a highly precise Attention U-Net architecture for automated MRI lesion localization, focusing on critical spatial features. |
TensorFlow/Keras, OpenCV |
| RL Autonomous Agents | Implemented and evaluated Q-Learning, SARSA, and Deep Q-Networks (DQN) for dynamic control in stochastic environments. | Python, OpenAI Gym |
- 📖 Book: Line Follower Robot: How to Build a Robot (2022)
- A 150-page technical foundation covering AVR microcontrollers, PCB design, and C-based embedded programming.
- 📄 Preprint: Performance Comparison of T5 and Marian MT Models (2023)
- Empirical evaluation of Neural Machine Translation models utilizing METEOR and BLEU metrics.
🌐 Web Development & Freelance Architecture (Click to expand)
While my primary focus is academic AI research, I possess a strong background in full-stack web development and server architecture. This elite engineering skill set allows me to seamlessly deploy complex deep learning models into production and manage robust IT infrastructures.
- Infrastructure Optimization: Secured and optimized high-traffic platforms including AI Lab TNB, Metadaru, and Kiyanet.
- Full-Stack Integration: Experienced in bridging the gap between heavy backend AI pipelines and interactive, intuitive web interfaces.