Md. Salauddin Sarker System Architect & AI Researcher

Data Science Researcher | Explainable AI (XAI) • Time-Series Forecasting • Industrial Analytics

Institutional Affiliation

Dept. of Educational Technology & Engineering (EdTE)

University of Frontier Technology (UFTB), Bangladesh

Researcher ID

ORCID: 0009-0005-7907-7522

I develop reliable and interpretable machine learning for real-world decision systems. My work combines explainable deep learning, time-series forecasting, and data pipelines to turn complex datasets into trustworthy insights for research and industry.

GPA: 3.84/4.00
Accepted (In Press) Publications
Industrial Data Engineering
Download CV (PDF) Publications Email
GitHub LinkedIn ORCID WhatsApp Telegram
Md. Salauddin Sarker - Professional Portrait

Two times

Dean's Awardee

Candidacy

Fall 2026 Ready

Scholarship Grade

Md. Salauddin Sarker - AI Research Context
"The complexity of industrial data requires the precision of neural interpretation."
Scientific Philosophy
Active Research Agenda

Scholarship
Aspiration.

Specializing in Trustworthy Data Science—architecting machine learning systems that maintain interpretability and robustness under real-world industrial constraints.

Explainable AI (XAI)

Developing faithful, actionable explanations for high-stakes decision makers using SHAP, LIME, and Grad-CAM.

Dynamic Forecasting

Evaluating neural resilience across non-stationary longitudinal data and multi-region normalizations.

4+ Accepted Papers
6+ Systems Deployed
VIEW FULL RESEARCH OUTPUT

Strategic AI Roadmap

Future Trajectories (2026-2030)

Knowledge-Grounded
LLMs

Integrating retrieval-augmented generation (RAG) with explainable reasoning to reduce hallucination in enterprise environments.

Target: Trustworthy AI

Recursive Agentic
Reasoning

Developing autonomous agents capable of self-correcting reasoning loops for complex multi-step industrial automation.

Target: Industrial Scale

Quantization-Aware
Edge AI

Optimizing deep models for low-power edge deployment in agri-tech and healthcare without sacrificing interpretability.

Target: Global Impact

Academic Pedigree

Educational Foundation

University of Frontier Technology

B.Sc. in Educational Technology & Engineering (EdTE)

University of Frontier Technology (UFTB), Bangladesh

CGPA: 3.84/4.00

2019 — 2024

Specialized Coursework

Artificial Intelligence Big Data Analytics Educational Robotics Advanced Statistics Industrial Automation
Dhaka College

Higher Secondary (Science)

Dhaka College

GPA: 5.00/5.00

SHKSC

Secondary School Certificate

SHKSC

GPA: 5.00/5.00

Research-Relevant Experience

Data Engineering & Operation Analytics

Q Cosmetics Logo

Software Developer / Data Systems

Q Cosmetics Ltd

May 2025 – Present

"Bridging the gap between industrial IoT data and enterprise intelligence. My work focuses on scalable ETL pipelines that serve as the foundation for trustworthy predictive analytics."

Operational Impact

  • → Rolled out multi-module ERP workflows, reducing order cycle time by 35%.
  • → Built Python–PostgreSQL ETL pipelines, improving audit consistency and data quality.

Analytics & BI

  • → Built KPI dashboards (Metabase/Tableau), reducing monthly reporting from 3 hours to 15 minutes.
  • → Structured master data and reconciliation logic for reliable downstream analytics.
Metrics are based on internal operational tracking and monthly reporting audits.

Technical Matrix.

Multi-Disciplinary Stack & Scientific Toolkit

Computational Stack
Python SQL Java PHP Dart Linux / Shell
ML / AI / XAI
PyTorch / TF SHAP / LIME Grad-CAM ARIMA / OLS Scikit-Learn Computer Vision
Data & BI
PostgreSQL Odoo (ERP) ETL Pipelines Tableau Metabase Data Warehousing
Proficiencies
Bengali Native
English Professional
Japanese Basic (N5)

Awards & Honors

Dean’s Award (2022, 2024)

University of Frontier Technology (UFTB)

Academic excellence for two non-consecutive cycles for top-tier GPA.

National STEAM Olympiad 2023

9th Place (National) • Campus Ambassador

Selected as campus leader to coordinate innovation drive; ranked top 10 nationally for technical pitch.

Academic Leadership

UFTB STEAM Club

Vice President

Facilitating technical mentorship for 200+ members and orchestrating high-fidelity engineering hackathons.

UFTB Robotics Club

Executive Member

Orchestrating large-scale robotics competitions and facilitating cross-discipline engineering hackathons.

UFTB Language Club

Japanese Language Secretary

Facilitating multi-lingual research communication and fostering global academic exchange.

Accepted
Scientific Output.

Registry of Accepted Peer-Reviewed Journals & Conferences

Springer Nature Q1 Journal Vol 2026

Leveraging Explainable AI for Sustainable Agriculture

Authors: Md. Salauddin Sarker, et al.

Scientific Methodology

Systematic mapping of 150+ papers; specialized 4-tier taxonomy for post-hoc local explanations in agri-systems.

Research Impact

Verified decision transparency across diverse field data; establishes baseline for trustworthy diagnostic deployment.

Impact Factor

Pending Final Release

REQUEST MANUSCRIPT
IEEE Xplore Full Paper ID: 273_COMPAS

Multi-Country Weather Forecasting with Hybrid ARIMA-OLS

Authors: Md. Salauddin Sarker, et al.

Architectural Approach

Hybridizing traditional statistics with modern regression on a 61-year longitudinal climate dataset.

Research Impact

Demonstrated 14.2% MAE reduction; verified neural resilience across severe non-stationary climate shifts.

Recognition

Accepted for IEEE COMPAS 2025

REQUEST MANUSCRIPT
IDAA 2025 Specialized Track

EdgeVision: Quantization-Aware Inference for Industrial Edge

Authors: Md. Salauddin Sarker, et al.

Optimization Focus

8-bit INT quantization on ARM platforms; Sub-10ms inference for low-power nodes.

Forum Status

Featured as High-Potential Research at the IDAA 2025 Scientific Forum.

REQUEST MANUSCRIPT
Ethics in AI Research Paper

Quantifying Algorithmic Fairness in Ed-Tech Systems

Authors: Md. Salauddin Sarker, et al.

Fairness Metrics

SHAP-based feature bias quantification; Demographic parity audit on success predictors.

Societal Impact

Establishes new benchmarks for fair Ed-Tech ML; Accepted at Ethics in AI Forum.

REQUEST MANUSCRIPT

Technical Validation.

Academic Certifications & Professional Recognition

IEEE COMPAS 2025 Certification
IEEE Certification

IEEE COMPAS 2025

IDAA 2025
Scientific Recognition

IDAA 2025: Featured Architect

Research Case Studies

Evidence of Technical Readiness & Scientific Method

Case Study #01 — Deep Learning + XAI

Rice Leaf Disease Detection

Building an accurate and interpretable classifier for sustainable agri-diagnostics.

The Problem

Need for reliable, transparent identification of leaf pathogens in field conditions.

The Data

Multi-class leaf image dataset; diverse environmental backgrounds.

Approach: Transfer learning (VGG16/ResNet50/EfficientNetB0) + attention mechanisms.

Explainability: Grad-CAM heatmaps for decision transparency and feature verification.

Results: 99.42% accuracy; sustained 42 FPS on mobile edge platforms.

RESEARCH IMPACT (MSc READY)

Demonstrated that spatial attention maps can pinpoint pathogenic regions in leaf images better than standard CNNs, crucial for trustworthy field diagnostic deployment.

REPOSITORY
Rice Leaf Disease XAI
Spatial Attention Heatmap Visualization
Case Study #02 — Time-Series

Multi-Region Weather Forecasting

Evaluating neural resilience across 61 years of non-stationary climatic data.

The Problem

Forecasting weather variables accurately across geographically diverse regions.

The Data

61 years of historical weather data across multiple regions.

Approach: ARIMA + OLS hybridization; comprehensive error diagnostics and baseline comparisons.

Metrics: Reduced MAE by 14.2% across validation sets; demonstrated 0.92 R-squared correlation.

KEY INSIGHT (LEARNING)

Identified that non-stationarity requires careful time-indexed evaluation and domain-aware baselines rather than just high-complexity architectures.

NOTEBOOK
Weather Forecasting
Multi-Region Variable Correlation
Case Study #04 — Industrial ML

Industrial Customer Churn Prediction

Bridging operational ERP data with predictive retention strategy.

The Problem

Predicting organizational churn and identifying key risk drivers in commercial flows.

The Output

85% prediction accuracy with feature importance insights for retention.

Approach: Classification models (XGBoost/RandomForest) + SHAP feature importance analysis.

Outcome: Actionable retention strategy insights for operational management.

REPOSITORY

Academic Endorsements

Farhana Islam

Farhana Islam

Assistant Professor & Chairman

Dept. of Educational Technology & Engineering

farhana0001@uftb.ac.bd

University of Frontier Technology (UFTB), Bangladesh

Aditya Rajbongshi

Aditya Rajbongshi

Assistant Professor

Dept. of Educational Technology & Engineering

aditya0001@uftb.ac.bd

University of Frontier Technology (UFTB), Bangladesh

Curriculum Vitae

Academic & Professional Summary

Research Direction: Seeking fully funded MSc (Thesis) opportunities in XAI + Time-Series + Reliable ML.

Data Systems: Proven industry experience in ERP Workflows + ETL Pipelines + BI Dashboards.

Scientific Output: Multiple Accepted Publications (In Press) across IEEE and Springer venues.

Download Full CV (PDF)

MSc
Candidacy 2026.

I am seeking a fully funded MSc with thesis (Fall 2026) to research Explainable AI, Time-Series Forecasting, and Trustworthy ML.

md.salauddin.sarker.icte@gmail.com