B.Sc. Electrical & Computer Engineering
2019 – 2024Rajshahi University of Engineering & Technology
Thesis: Ensemble of Convolutional Neural Networks for Stroke Prediction
Md Shahriar Sajid
Lecturer, Department of Computer Science & Engineering
European University of Bangladesh · Research Assistant, Transparent Intelligence Lab (RUET)
Decoding neural signals
- Based
- Dhaka, Bangladesh
- Degree
- B.Sc. Electrical & Computer Engineering, RUET
- English
- IELTS 7.5 · CEFR C1
- Focus
- EEG · BCI · Privacy-Preserving ML · Computer Vision
Actively looking for PhD and Master's opportunities
01
About
I am a Lecturer in Computer Science and Engineering at the European University of Bangladesh. I also work as a Research Assistant at the Transparent Intelligence Lab at RUET, with Dr. Md Nahiduzzaman, and at the ELITE Research Lab, with Md Kishor Morol. I hold a B.Sc. in Electrical & Computer Engineering from Rajshahi University of Engineering & Technology.
My work sits at the intersection of machine learning and biomedical signal processing. I design lightweight, interpretable architectures for EEG analysis particularly brain-computer interfaces, and study how well they generalize to unseen subjects. I have also worked on privacy-preserving federated learning for decentralized healthcare and on computer vision for remote sensing. A recurring result across these projects is that simpler methods often generalize better when data is scarce: my current work on cross-subject EEG locomotion decoding shows classical spatial filtering outperforming deep models under artifact-aware, subject-wise evaluation. Long term, I want to contribute to clinical AI systems reliable enough to be used outside controlled experimental settings.
02
Research Interests
- Machine learning & deep learning
- Biomedical signal processing
- Brain–computer interfaces
- Cross-subject generalization
- Privacy-preserving & federated learning
- Computer vision & remote sensing
- Interpretability & trustworthy AI
- Time-series analysis
03
Education
IELTS Academic — Overall 7.5
Aug 2025British Council / IDP / Cambridge · CEFR C1
- Listening 8.0
- Reading 7.5
- Writing 7.0
- Speaking 7.0
04
Publications
2 peer-reviewed · 2 in preparation
PaperNet: Efficient Temporal Convolutions and Channel Residual Attention for EEG Epilepsy Detection
Intl. Conf. on Intelligent Data Analysis and Applications (IDAA)
DOI ↗Privacy Preserving Machine Learning Model Personalization through Federated Personalized Learning
Intl. Conf. on Data Analytics for Business and Industry (ICDABI), Bahrain
DOI ↗Artifact Removal in Cross-Subject EEG Locomotion Decoding: A Comparison of Classical and Deep Learning Models
Benchmarking Deep Learning and Vision Transformers for Multi-Platform Land Cover Scene Classification in Bangladesh
05
Selected Research Projects
EEG · 2025
PaperNet — Lightweight EEG Epilepsy Detection
Temporal convolutions, channel-wise residual attention, and a bidirectional LSTM for short-window EEG classification.
- Params~0.6 M
- Macro-F10.958
- ROC-AUC0.997
- Accuracy0.958
BEED dataset · TensorFlow / Keras
EEG · Ongoing
Cross-Subject EEG Locomotion Decoding
Classical vs. deep pipelines for binary rest/walk decoding under EEG+EOG and EEG-only settings, with subject-wise GroupKFold.
- Mean Acc.79.3%
- Macro-F10.792
- Std. Acc.4.6%
- BestCSP + SVC
OpenNeuro ds007526 · CSP, RF, XGBoost, EEGNet
FL · 2023
Privacy-Preserving Federated Personalization
Benchmarking federated personalization against differential privacy, homomorphic encryption, secure aggregation, and SMPC.
- BestAPPLE+HE
- Accuracy99.34%
- F1-Score99.34%
- Clients200
VirusMNIST · Published at ICDABI 2023
Clinical AI · Thesis
Ensemble CNNs for Stroke Prediction
An ensemble of convolutional neural networks developed to improve diagnostic accuracy in stroke prediction.
- DomainClinical AI
- ApproachCNN Ensemble
RUET · Dept. of Electrical & Computer Engineering
Vision · Ongoing
Land Cover Classification in Bangladesh
A benchmark of CNNs and Vision Transformers on the MLC-BD dataset for multi-platform land cover scene classification.
- ModelsCNN · ViT
- DomainRemote Sensing
Working paper · Geospatial computer vision
Open Source · 2024
aosPB — AOSP-Based Android Firmware
Founder and lead developer of a Vanilla AOSP-based custom Android firmware built for multiple hardware platforms.
- SinceOct 2024
- StackC · Bash · YML
Open source · CI/CD via GitHub Actions
06
Experience
Research Assistant — ELITE Research Lab
2026 – PresentPI: Md Kishor Morol, CEO, ELITE Research Lab, NY, USA
Applied machine learning research with an international team.
Research Assistant — Transparent Intelligence Lab
2025 – PresentPI: Md Nahiduzzaman, Asst. Prof., Dept. of ECE, RUET
EEG brain–computer interfaces, cross-subject generalization, interpretable biomedical ML.
Lecturer — Dept. of Computer Science & Engineering
2025 – PresentEuropean University of Bangladesh, Dhaka
Teaching programming and applied ML; supervising undergraduate theses.
Teaching & Service
Courses
Computer Programming · Applied Machine Learning
Supervision
Real-Time Traffic Sign Detection Using CNN
Hasan, Hossain, Jannat-E-Nur, Riya
Hasan, Hossain, Jannat-E-Nur, Riya
Impacts of COVID-19 Vaccines among Bangladeshi Adults
Azme, Ullah
Azme, Ullah
Mentoring
Chief Mentor, Programming Club, European University of Bangladesh — 50+ students.
Open source
Founder & Lead Developer of aosPB, an AOSP-based custom
Android firmware.
07
Skills & Awards
Technical Skills
Languages
Python · C · C++ · Dart · Bash · MySQL
ML / DL
PyTorch · TensorFlow · Keras · Scikit-learn · OpenCV
Tools
Git · GitHub Actions · CircleCI · Drone CI · Linux · LaTeX · AOSP
Other
Technical writing · Reproducible research · Project management
Awards & Recognition
Stipend
Merit-Based University Stipend — received annually throughout
the undergraduate program at RUET.
2023
International University Physics Competition — accomplished
competitor.
2021–25
Hacktoberfest — open source contributor
sajidshahriar72543
08
Contact
I am open to research collaborations, graduate study opportunities, and conversations about EEG decoding, biomedical machine learning, and privacy-preserving AI.
- GitHub
- @sajidshahriar72543
- Location
- Dhaka, Bangladesh