Computer Science Researcher & Systems-Minded Builder
Daffodil International University
Department of Computer Science & Engineering
Dhaka, Bangladesh · 23.8103° N / 90.4125° E
Research philosophy, foundational principles, and cross-disciplinary trajectory.
Md. Tanvir Hasan is a computer science engineering researcher and systems architect working at the intersection of Deep Learning, Explainable AI (XAI), Medical Image Analysis, and Distributed Compute Infrastructure.
His investigations focus on solving high-stakes challenges where algorithmic opacity carries real-world consequences: developing interpretable hybrid CNN-Vision Transformer frameworks for clinical screening, knowledge distillation for resource-constrained edge devices, and carbon-aware GPU scheduling pipelines for high-performance computing clusters.
With 16+ peer-reviewed book chapters, conference proceedings, open research datasets, and manuscripts under review, his work couples rigorous empirical methodology with practical, open-source software engineering. Alongside his academic research, he leads distributed AI infrastructure platform engineering at Daffodil International University (ClusterOne DevTeam).
Core theoretical concentrations and active research directions.
Segmentation and classification pipelines on Ultrasound, MRI (multi-sequence), Histopathological Pap-smear, and Dental Panoramic Radiographs. Focus on robust multi-clinic generalization and high-resolution diagnostic feature extraction.
Calibrating uncertainty and interpreting black-box neural representations in critical decision pipelines. Grounding diagnostic predictions through Grad-CAM, Score-CAM, SHAP, and cross-modal evidential deep learning.
Transferring rich representation capacity from massive multi-modal ensembles into lightweight, low-latency student models for edge deployment in resource-limited clinical environments.
High-performance orchestration of campus-scale heterogeneous GPU compute pools using Kubernetes and Slurm. Developing adaptive reinforcement learning models (NeuralSched) for carbon-aware job placement and failure recovery.
Formal language text and syntax security parsers using Lex & Bison, controlled endpoint USB hardware injection simulations, and quantum-resilient post-quantum cryptographic schemes for distributed ledgers.
Multi-disciplinary academic trajectory spanning computer science and foundational studies.
Daffodil International University (DIU), Dhaka, Bangladesh
Cumulative GPA: 3.?? / 4.00
Key Coursework: Algorithms & Data Structures, Artificial Intelligence, Machine Learning, Operating Systems, Computer Architecture, Compiler Design, Database Management Systems, Distributed Computing, Cryptography & Network Security.
Islamic Arabic University, Dhaka
Status: Active / Ongoing
Advanced classical methodology, foundational jurisprudence, and canonical textual criticism.
Islamic Arabic University, Dhaka
Result: 3.83 / 5.00 (First Class Honors)
Ta'mirul Millat Kamil Madrasah, Tongi, Gazipur
Result: GPA 5.00 / 5.00 (Golden / Perfect Score)
General Science & Theoretical Studies.
16 peer-reviewed book chapters, conference papers, open datasets, and research manuscripts with BibTeX citations.
Intelligent Data Analytics and Applications (IDAA 2025) · Taylor & Francis Book Chapter, 2025.
International Journal of Computer Applications (IJCA), Vol. 187, No. 100, Foundation of Computer Science (FCS), NY, USA, 2024.
2025 28th International Conference on Computer and Information Technology (ICCIT) · IEEE, 2025.
2026 International Conference on Artificial Intelligence and Sustainable Energy Innovations (AISEI) · IEEE, 2026.
Mendeley Data Repository · DOI: 10.17632/t66ytdrhf2.1, 2026.
Data Paper Manuscript · Companion descriptor for the MOPG-7 benchmark, 2026.
Accepted for publication · Forthcoming, 2026.
Accepted for publication · Forthcoming, 2026.
Under review · Target: Q1 Medical Imaging Journal, 2026.
Submitted · Under review, 2026.
Submitted · Under review, 2026.
Submitted · Under review, 2026.
Submitted · Under review, 2026.
Ongoing scientific investigation, 2026.
Ongoing scientific investigation, 2026.
Ongoing scientific investigation, 2026.
Engineering practical infrastructure, compute pipelines, and academic community leadership.
Daffodil International University · Distributed Compute Division
Engineered campus-wide distributed AI infrastructure platform (ClusterOne) uniting pooled GPU accelerators (A100/V100/RTX series) for concurrent departmental deep learning research.
DIU Research Society (DIURS)
Mentored undergraduate researchers on scientific writing, reproducible experimental methodologies, LaTeX formatting, and literature review synthesis.
Architected tools, research experiments, and software systems designed to solve difficult problems.
Hackathon Champion (1st Place among 100+ competing teams). Adaptive deep reinforcement learning orchestrator for predictive GPU failure detection and carbon-aware batch job scheduling across heterogeneous clusters. Features automated checkpoint migration prior to thermal throttling events.
Deterministic lexical parser and static analyzer engineered in Lex and Bison for identifying suspicious token sequences, unvalidated memory operations, and injection hazards prior to runtime execution.
Hardware-software endpoint security simulation verifying defensive logging efficacy against high-speed USB Human Interface Device (HID) keystroke injection attacks under isolated lab constraints.
Explainable AI-powered fraud intelligence, scam protection, money-mule detection, and disaster-aware agent liquidity resilience engineered for Mobile Financial Services (MFS). Secured 5th place out of 132 university teams nationwide at upay presents Inter University AI DEV Fest 2026.
Geospatial comparative analysis system developed for the NASA Space Apps Challenge 2026 (Dhaka, Bangladesh). Evaluated terrestrial planetary analogs correlating geological, thermal, and isolation metrics with prospective permanent lunar surface bases and martian habitats.
Industry-verified credentials validating competencies in cloud architectures and defensive cybersecurity.
Amazon Web Services (AWS) Academy
Verify Credly Badge ↗Scientific community initiatives, research competitions, and academic volunteer service.
Secured 1st Place / Champion out of 100+ competing teams with NeuralSched (neuralsched.vercel.app ↗) — Adaptive deep reinforcement learning for predictive failure detection and carbon-aware distributed GPU job routing.
National AI & MFS Hackathon organized by DIU CPC × upay (UCB Fintech Company Limited). Secured 5th position among 132 competing university teams nationwide with TakaSafe (takasafe-diu.surge.sh ↗).
Global space & science hackathon organized by NASA. Tackled challenge: Identify Earth Locations that Analog the Permanent Moon Base Locations and Mars with project ShadowAstra Atlas (shadowastra-atlas.vercel.app ↗).
Division of Research (DoR), Daffodil International University.
Army IBA Sylhet · Focus on SDG-Driven Artificial Intelligence & Healthcare Research.
Organizing team, coordination of high-school and university machine learning problem solvers.
Student research coordination, peer review circles, and seminar organization.
Competitive algorithmic problem solving, graph theory, and dynamic programming challenges.
Languages, deep learning frameworks, cluster tools, and research environments.
Available for scientific collaborations, graduate research assistantships, and systems engineering roles.
I am actively seeking funded Ph.D. / M.S. graduate research assistantships in Computer Science, specializing in Medical Computer Vision, Explainable AI, and High-Performance Compute Systems.