Professional Experience
Research Experience
Remote Research Assistant
Washington State University — School of Electrical Engineering and Computer Science
June 2026 – Present
Project: Anomaly-Aware ForensiBlock · Supervised by Prof. Asma Jodeiri Akbarfam
- Contributing to the AI/ML and explainable AI pipeline for a real-time authorized-user anomaly detection system
- Developing anomaly detection and XAI integration for interpretable decision support and forensic traceability, under external faculty supervision
Independent research lab working on efficient, explainable deep learning for high-stakes decisions in medical imaging, biosignals, and trustworthy AI.
- Founded the lab in 2026 and lead its research agenda on efficient, explainable deep learning for clinical applications in medical imaging and intelligent health systems
- Lead or co-author on 9 manuscripts spanning computer vision, signal modelling, and trustworthy AI — 3 accepted at IEEE BECITHCON 2026 and iCONEECT 2026, 5 under review (IEEE ICCIT 2026, IEEE JBHI, IEEE Access, and PLOS ONE), and 1 manuscript in preparation
- Directing work on lightweight CNN-Transformer architectures, attention and gating fusion, structural reparameterization, imbalanced learning, and explainability methods (Grad-CAM, SHAP) under subject-independent evaluation protocols
- Established lab infrastructure, reproducible research pipelines, and collaboration frameworks with researchers at ELITE Research Lab, Washington State University, Merit Network / University of Michigan, and the Medical University of Vienna
Elite Research Lab
December 2025 – Present · 3 roles
Research Assistant
September 2026 – PresentProject: LCM-UNet — Reparameterizable Local-Compensated Mamba U-Net · Joint First Author
- Joint first-author work on LCM-UNet, a 34K-parameter Mamba-based skin lesion segmentation network; developed and evaluated the Residual Dilated Local Adapter (RDLA) for improved local boundary modeling with no additional inference-time operator after reparameterization
- Achieved +0.48/+0.37 percentage-point IoU gains on ISIC 2017/2018 and 86.37 ± 0.29% IoU on PH²; conducted multi-seed, boundary-sensitive, statistical, and fusion-equivalence evaluation while retaining 34,180 parameters and 0.05974 GFLOPs
Junior Research Assistant
March 2026 – August 2026Project: BGD-SF PolySegNet — Boundary-Guided Polyp Segmentation · First Author
- First-author work on BGD-SF PolySegNet, developing boundary-guided multi-scale aggregation, selective skip fusion, and multi-level boundary refinement for colonoscopy polyp segmentation
- Built a multi-source pipeline using 3,535 images from three datasets and evaluated generalization on four external datasets; achieved 0.887 Dice, 0.839 IoU, and 0.531 Boundary F1 on the internal test set
- Improved Boundary F1 by 15.7% and reduced HD95 by 27.5% and ASSD by 34.6% versus the strongest corresponding CNN baselines; performed ablation and statistical significance analysis
Research Intern
December 2025 – February 2026- Supported medical imaging research through dataset preprocessing, baseline model reproduction, experiment execution, and quantitative result comparison
- Assisted with literature review, experiment tracking, result tables/figures, and technical documentation for ongoing deep learning research projects
Undergraduate Thesis Researcher
AUST & Monash University (Fellow Institute), Dhaka, Bangladesh
2023 - 2024
Project: Video-Based Driver Distraction Detection: Bangladesh Perspective
- Collaborated with a Monash University PhD researcher and applied transfer learning across CNN architectures with dataset curation and cross-architecture benchmarking
- Co-authored a peer-reviewed paper published at IEEE iCACCESS 2024
Industry Experience
Software Engineer I
Brain Station 23 PLC
March 2025 - Present
OneID – Backend Engineer
Tech Stack: FastAPI PostgreSQL (JSONB) Keycloak
- Engineered an eKYC data-processing pipeline that reduced third-party integration latency by 40%
- Increased document processing speed by 35% using JSONB metadata system
- Reduced invalid submissions by 30% with NID validation and normalization
- Secured sensitive identity data using Keycloak encryption
- Enhanced search and filtering performance by 25% for faster data retrieval
OneID – Client Portal (Django Admin)
Tech Stack: Django Keycloak
- Improved admin efficiency by 25% with a real-time analytics dashboard
- Reduced document errors by 20% via improved upload and status flows
- Secured user sessions and optimized logout with Keycloak compliance
ObhaiVTS – Frontend Engineer
Tech Stack: React Tailwind Shadcn Zustand TanStack Query
- Boosted data handling by 30% with server-side pagination and dynamic filtering
- Improved app responsiveness using TanStack Query caching and real-time updates
- Reduced UI glitches by 20% with Zustand for centralized state management
- Delivered Invoice, Customer, Vehicle & Hub management flows with smooth UX
Teaching Experience
Adjunct Lecturer, CSE Department
International Standard University (ISU)
January 2026 - Present
- Taught Robotics, IoT, and Algorithms to undergraduate students, aligning coursework with industry standards
- Prepared students for real-world projects and competitions, improving practical skills and employability
- Designed hands-on labs and problem-solving sessions to strengthen technical understanding
Competitive Programming Mentor (Remote)
Shohoj Coding
September 2024 - December 2025
- Guided 50+ students in Graph Theory, Number Theory, and Data Structures, improving problem-solving skills
- Boosted student contest performance by 30% through targeted mentoring and practice sessions
- Developed structured learning modules to simplify complex algorithms for beginners