$ ssh researcher@mdmehedihasan.us
Authenticating... OK
Loading research_profile.json [10 papers found]
Mounting /projects, /publications, /awards...
Welcome, guest. Session ready
  whoami.sh mehedi@research:~$
mehedi@research:~$whoami

// mission
Building interpretable machine learning systems at the intersection of behavioral modeling, social engineering detection, and socio-technical resilience.

./view_research.sh ↓ download_cv.pdf orcid ↗
3.76CGPA / 4.00
10+Accepted Papers
🏆NASA Apps · 3 Honors
24Research Works
Md. Mehedi Hasan — AI & ML Researcher
AI/ML Research
SEM Modeling
President DIU CPC
Explainable AI
Behavioral Modeling
$ cat about.md

Turning Behavior Into Intelligence

about.md

## Who I Am

I work at the intersection of AI, behavioral modeling, and socio-technical resilience — building systems that are not only accurate but explainable and grounded in human reality.

I'm a BSc Computer Science & Engineering student at Dhaka International University (CGPA 3.76 / 4.00), graduating December 2027. My research focuses on detecting social engineering through interpretable behavioral anomaly modeling, predicting system failures using socio-economic and IoT data, and building continual learning frameworks for dynamic environments.

As President of DIU Computer Programming Club and an active IEEE & IEEE Computer Society Member, I combine research depth with community leadership — recognized as a Regional Winner (Barisal Region), Global Nominee, and Honorable Mention for the Global Finalist at NASA International Space Apps Challenge 2025 with Team Polaris.

IEEE & IEEE CS Student Member Behavioral ML Social Engineering Detection Interpretable AI Structural Equation Modeling (SEM) Loss Functions (InfoNCE, Focal) Continual Learning Geospatial Risk Computer Vision & YOLOv8 LLMs & GenAI

fig.0 — research_methodology_framework.svg

BEHAVIOR /SIGNAL DATA INTERPRET.ML LAYER FRAUD / RISK RETENTION GEO / ORBITAL
research_formulations_sem_models.py

# Core Mathematical Formulations & Architectures

[1] Loss.InfoNCE (MEWC-RL++)
L = -log(e^(z_i·z_j/τ) / ∑ e^(z_i·z_k/τ))
[2] SHAP Attribution
ϕ_i(v) = ∑ [|S|!(|N|-|S|-1)!/|N|!] Δv_i(S)
[3] SEM Structural Model
η = Bη + Γξ + ζ (CFI=0.982, RMSEA=0.031)
[4] DistilBERT & YOLOv8
Softmax(QK^T/√d_k)V || FPN C2f Head

✓ Formulations underlying 10+ accepted papers in behavioral ML, social engineering detection, and AI resilience.

$ cat education.log
Dhaka International University
B.Sc. Computer Science & Engineering
Sep 2023 – Dec 2027 · CGPA 3.76 / 4.00
BAF Shaheen College Jashore
Higher Secondary Certificate (HSC)
Sep 2020 – Dec 2022 · GPA 4.75 / 5.00
$ stat --profile
3.76CGPA
10+Accepted
12Under Review
3Ongoing
$ cat affiliations.txt
IEEE — Student Member (thru Dec 2026)
IEEE Computer Society — Student Member
BASIS — Student Member
$ ls publications/ --sort=status

Publications

Accepted, under review, and ongoing research across ML, AI, socio-economics, and computer vision.

01
acceptedBIM 2025 · Taylor & Francis

A Socio-Economic Machine Learning Framework for Predicting Programmer Retention

Hasan, M. M., Rakib, R., Molla, M. A., Borhan, R., Based, M. A.

02
acceptedBIM 2025 · Taylor & Francis

Fraud Detection with DistilBERT: A Transformer-Based Approach to Behavioral Banking Sequences

Hasan, M. M., Mahin, A. A., Chakraborty, S., Afrose, M., Mia, M. A., Based, M. A.

03
acceptedECCT 2026 · Taylor & Francis

MEWC-RL++: A Modular and Adaptive Continual Reinforcement Learning Framework for Automation Systems

Hasan, M. M., Mahin, A. A., Afrose, M., Based, M. A.

04
acceptedECCT 2026 · Taylor & Francis

Impact of Social Media Usage on Academic Performance: A Hybrid Machine Learning and SEM Approach

Mrida, M. A. R., Hasan, M. M., Ausaf, S. M. N., Ray, P. C., Based, M. A.

05
acceptedECCT 2026 · ACSR, Springer Nature

Lightweight Deep Learning for Urban Traffic Intelligence: A MobileNetV2-YOLOv8 Pipeline with Adaptive Resolution Processing

Biswas, S., Hasan, M. M., Rakib, R., Mahin, A. A., Afrose, M., Based, M. A.

06
acceptedECCT 2026 · Taylor & Francis

Federated and Explainable Machine Learning Framework for Cyber Attack Detection in Smart Grids

Asif, M., Hasan, M. M., Based, M. A.

07
acceptedIUB Journal of Science and Engineering (AJSE)

Student-YOLOv8: A YOLO-Based Deep Learning Approach for Real-Time Student Detection and Face Recognition in Classroom Attendance Monitoring Systems

Hasan, M. M., Hossain, M. A., Mahin, A. A., Khatun, F.

08
acceptedTENCON 2026-IEEE

Cross-Cultural Early Detection of Suicidal Ideation in Adolescents: An Ensemble Machine Learning Approach Using GSHS Data

Rion, A. M., Pallob, M. M. I., Rakib, R., Molla, M. A., Hasan, M. M.

09
acceptedMIET 2026 · Springer Nature LNNS

A Hybrid Explainable AI Framework for Concurrent Earthquake Magnitude Estimation and Severity Classification

Molla, M. A., Rakib, R., Hasan, M. M., Pallob, M. I., Rion, A. M., Based, M. A.

10
acceptedICAIMS 2026-Malaysia-IEEE

FairCF: Fair Counterfactual Explanations for Student Academic Performance Prediction

Molla, M. A., Rakib, R., Pallob, M. I., Hasan, M. M., Rion, A. M.

R1
under reviewJournal

Digital Transformation for Sustainable Universities: Modeling the Convergence of MIS, IoT, and Governance

Hasan, M. M., Pallob, M. M. I., Based, M. A.

R2
under reviewJournal

War, Media Bias, and Socio-Economic Consequences: A Centennial Analysis of the Narrative-Reality Index (1914–2026)

Hasan, M. M., Mahin, A. A., Afrose, M., Based, M. A.

R3
under review

A Hybrid Explainable Machine Learning Framework for Customer Lifetime Value Prediction and Its Socio-Economic Implications

Hasan, M. M., Rahman, M. M., Mrida, M. A. R., Based, M. A.

R4
under review

A Relational Database-Driven Real-Time Cryptocurrency Portfolio Management System with Automated ETL Integration

Rakib, R., Hasan, M. M., Molla, M. A., Based, M. A.

R5
under review

SafeRoads: AI-Driven Proactive Risk Synthesis and Data-Centric Infrastructure Audits for Road Safety

Rahman, M. M., Hasan, M. M., Based, M. A.

R6
under review

Machine Learning based Regression and Classification of Earthquake Magnitude using USGS Seismic Records

Molla, M. A., Rakib, R., Hasan, M. M., Rion, A. M., Based, M. A.

R7
under review

Urban–Rural Inequality in Tech Career Intention: A Machine Learning Perspective

Molla, M. A., Hasan, M. M., Rakib, R., Borhan, R., Based, M. A.

R8
under review

Socio-Economic Determinants of Programming Skill Development in Undergraduate CS Students

Hasan, M. M., Molla, M. A., Rakib, R., Borhan, R., Based, M. A.

R9
under review

Explainable AI for Understanding Motivation in Choosing Computer Science Careers

Rakib, R., Molla, M. A., Hasan, M. M., Borhan, R., Based, M. A.

R10
under review

Fairness-Aware HIV Risk Prediction in Bangladesh Using Behavioral and Explainable AI

Afrose, M., Hasan, M. M., Mahin, A. A., Based, M. A.

R11
under review

A Modular Framework for Continual Reinforcement Learning in Dynamic Robotic Environments

Mahin, A. A., Hasan, M. M., Rakib, R., Molla, M. A., Based, M. A.

R12
under review

Interpretable Machine Learning for Depression Risk Classification Among Bangladeshi University Students

Mahin, A. A., Hasan, M. M., Afrose, M., Mia, M. A., Based, M. A.

O1
ongoing

Women's Psychology and Its Socioeconomic Impact on Male Partners and Relationships

Hasan, M. M., Rakib, R., Molla, M. A., Based, M. A.

O2
ongoing

THE DUAL PARADOX: AI Investment and Workforce Displacement in Global Corporations

Hasan, M. M., Molla, M. A., Rakib, R., Based, M. A.

O3
ongoing

MEWC-LLM: Memory-Efficient Weight Consolidation for Large Language Models

Afrose, M., Hasan, M. M., Based, M. A.

$ ls -la ~/projects

Projects

Research prototypes and engineering systems spanning ML, CV, and full-stack development.

↗
📖 handbook

The AI Researcher's Handbook

A comprehensive guide for AI/ML researchers covering research methodology, paper writing, experimentation design, and navigating the academic publication process. Available on ResearchGate.

ResearchGateOpen AccessMethodology

fig.4 — academic_lifecycle.svg

HYPOTHESIS EXPERIMENT DRAFT REVIEW
↗
🏆 nasa_space_apps · 3 honors
🌍

MeteorShield

NASA NEO data platform with real-time 3D visualization. Regional Winner (Barisal Region), Global Nominee, and Honorable Mention for the Global Finalist at NASA International Space Apps Challenge 2025 with Team Polaris.

NASA APIThree.jsMachine Learning3D Vis

fig.3 — meteorshield_pipeline.svg

API 3D VIS DASHBOARD
↗
🤖

SocialAI

Distributed, scalable platform integrating LLM agents and resilient proxy routing for automated content generation.

LLMsPythonAutomation

fig.5 — distributed_llm_pipeline.svg

LLM AGENTS PROXY POOL TARGET API
↗
🍃

SaveFood

XGBoost model trained on IoT time-series data to predict food spoilage (F1: 0.89). Dashboard visualizes waste reduction impact.

XGBoostIoT

fig.6 — spoilage_prediction_model.svg

TEMP/HUMID GAS SENSORS XGBOOST
↗
⚡

DevBurn

TypeScript full-stack engine analyzing developer workflows, integrating GitHub services with burnout tracking algorithms.

TypeScriptAnalytics

fig.7 — burnout_telemetry.svg

GITHUBEVENTS TRACKINGALGORITHM
↗
🛣️

SafeRoads

ML model identifying high-risk urban segments using traffic patterns and accident history (Precision: 84%).

Geospatial MLEnsemble

fig.2 — risk_synthesis.svg

TRAFFIC HISTORY RISKMODEL
↗
📡

HTSPF

Hierarchical Temporal-Spatial Pattern Fusion model for advanced data analysis and forecasting in dynamic environments.

Deep LearningSpatial-Temporal

fig.1 — hierarchical_fusion.svg

TIME IN SPACE IN HIER.FUSION
↗
🌐 LIVE DEMO · meetmehedi.github.io/Tactic-Detector
🕵️

Tactic Detector

Social Engineering Tactic Detection system using DistilBERT + SHAP attribution to identify and explain psychological manipulation tactics in text. MIT Licensed · 1★ on GitHub.

PythonDistilBERTSHAP XAINLP

fig.9 — tactic_detection_pipeline.svg

TEXT INPUT Distil BERT SHAP EXPLAIN SE
↗
🟢 NEW · 2026
🏥

SmartHealthAI

AI-powered smart health monitoring dashboard with real-time vitals tracking, predictive analytics, and ML-driven health risk assessment. Live demo on GitHub Pages.

HTML/CSS/JSMLHealthcare AI

fig.10 — health_monitoring.svg

VITALS SENSORS RISK AI
↗
🌐 LIVE · murnitin.vercel.app
🧩

murnitin

Python-based intelligent system with live deployment on Vercel. 2★ on GitHub. Incorporates ML pipelines and an API-driven backend for real-world research-grade inference.

PythonML PipelineVercel

fig.11 — inference_pipeline.svg

API IN ML PIPELINE OUT
↗
🧠

AI Learning Platform

AI-powered supplementary educational platform enhancing university coursework with intelligent assistance and interactive tracking.

TypeScriptEdTechAI

fig.8 — educational_assistant_loop.svg

SYLLABUS IN STUDENT LOG AI ASSIST
↗
₿

CryptoTracker

Real-time cryptocurrency portfolio management application for tracking live market data and analyzing asset performance.

WebAPIFinance

fig.9 — market_api_stream.svg

LIVE API PORTFOLIOTRACKER
HTSPS mewcLLM mewcRL Tactic Detector SEM Correlation DistilBERT
$ ./list_skills --category=all

Skills & Technical Ecosystem

Novel AI architectures created, programming languages, ML/DL frameworks, and engineering tools.

⚡ CREATED FRAMEWORKS

Novel Architectures & Formulations

Original AI models & empirical frameworks designed in research publications:

MEWC-RL++ (InfoNCE) HTSPF (Spatial-Temporal) Student-YOLOv8 Distillation Tactic Detector (SHAP XAI) MEWC-LLM Proxy Engine Socio-Econ PLS-SEM Path
💻 LANGUAGES

Programming Languages

Core languages used for research pipelines, ML modeling, and production code:

Python 3.12+ C / C++ TypeScript JavaScript (ESNext) SQL HTML5 / Vanilla CSS R LaTeX
🧠 LIBRARIES & ML

AI / ML / DL Libraries

Core scientific compute & deep learning stacks:

PyTorch TensorFlow Scikit-learn HuggingFace Transformers YOLOv8 (Ultralytics) OpenCV SHAP (Explainable AI) XGBoost SciPy & Statsmodels PLS-PM
🛠️ TOOLS & INFRA

Engineering & Tools

Development environments, automation, & infrastructure:

Git & GitHub Actions Docker Vercel Deployment Linux / Bash Shell N8N Workflow Automation Three.js 3D PostgreSQL Google Cloud / Colab Selenium / Playwright
MEWC-RL++HTSPFStudent-YOLOv8Tactic DetectorPLS-SEM PathInfoNCE LossmewcLLMDistilBERTSHAP XAIXGBoostPyTorchTensorFlowComputer VisionNLP
$ cat awards.log leadership.log

Awards & Leadership

🏆
NASA International Space Apps Challenge 2025

Regional Winner · Global Nominee · Honorable Mention — Global Finalist

Led Team Polaris to Regional Winner at the Barisal Region, then Global Nominee and Honorable Mention for the Global Finalist at NASA International Space Apps Challenge 2025. Combined expertise in Machine Learning, full-stack development, data analysis, and UI/UX to engineer a winning planetary defense solution — MeteorShield.

Machine Learning3D VisualizationNASA APITeam Polaris
leadership roles
  • President
    DIU Computer Programming Club · Apr 2026 – Present
  • General Secretary
    DIU CSE Speakers Club · May 2025 – Mar 2026
  • Head, Research & Training Wing
    DIU Career Development Club · Aug – Dec 2025
  • Executive Member
    BASIS Students' Forum (DIU) · Sep 2024 – Oct 2025
open for research collaborations
$ cat achievements.log
NASA International Space Apps Regional Champion, Global Nominee and Honorable mention at global finalist
National Data Analytics Competition 2025Participant
Eastern Bank PLC Technovation'25National Hackathon · Participant
BUET CSE FEST 2026National Hackathon · Participant
$ ./contact --open

Let's Shape
the Future

Open for research collaborations, engineering challenges, or just a friendly conversation about AI and its impact on humanity.

meetmehedi1@gmail.com
send_message.sh

Opens your email client with the message pre-filled

● STATUS: OPEN FOR RESEARCH COLLABORATION PUBLICATIONS: 10 ACCEPTED · 12 UNDER REVIEW · 3 ONGOING 🏆 NASA SPACE APPS 2025 — REGIONAL WINNER · GLOBAL NOMINEE · HONORABLE MENTION PROFILE: AI & ML RESEARCHER · INTERPRETABLE ML CORE FORMULATIONS: InfoNCE LOSS · SEM PATH MODELING · SHAP ATTRIBUTION · DistilBERT CGPA: 3.76 / 4.00 · DHAKA INTERNATIONAL UNIVERSITY PRESIDENT, DIU COMPUTER PROGRAMMING CLUB IEEE & IEEE COMPUTER SOCIETY STUDENT MEMBER ● STATUS: OPEN FOR RESEARCH COLLABORATION PUBLICATIONS: 10 ACCEPTED · 12 UNDER REVIEW · 3 ONGOING 🏆 NASA SPACE APPS 2025 — REGIONAL WINNER · GLOBAL NOMINEE · HONORABLE MENTION PROFILE: AI & ML RESEARCHER · INTERPRETABLE ML CGPA: 3.76 / 4.00 · DHAKA INTERNATIONAL UNIVERSITY
↑
$ cat publication_abstract.txt

// RESEARCH ABSTRACT
// BIBTEX ENTRY