Skip to main content
Fall 2026 MSc Candidate UFTB · EdTE

Md. Salauddin
Sarker

System Architect & AI Researcher

Researching

Building reliable, interpretable ML for real-world decisions — with a focus on Explainable AI (XAI) and Time-Series Forecasting.

Affiliation

Dept. of EdTE — UFTB, Bangladesh

  • GPA 3.84 / 4.00
  • Output Published Research
  • Domain Industrial Analytics
Md. Salauddin Sarker - Professional Portrait

Open To

Thesis-Track Supervision

XAI · Time-Series

Two Times

Dean's Awardee

Academic Distinction

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+ Published 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 out of 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 out of 5.00

SHKSC

Secondary School Certificate

SHKSC

GPA: 5.00 out of 5.00

Research-Relevant Experience

Data Engineering & Operation Analytics

Q Cosmetics Logo

Software Developer / Data Systems

Q Cosmetics Ltd

January 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 (IELTS: 6.5) 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.

Research Profile

Research
Publications.

Selected journal, conference, and manuscript outputs arranged for quick judgment on venue quality, indexing strength, and research fit.

For Prospective Supervisors

Built to be understood in one supervisor scan.

Venue strength, indexing, and research fit are surfaced first. Methods and abstract stay secondary.

Best venue

Q1 journal in Artificial Intelligence

Artificial Intelligence Review under Springer Nature

Indexed reach

SCIE, Scopus, and IEEE Xplore exposure

Journal article, conference proceeding, and manuscript-stage work

Research fit

XAI, forecasting, edge AI, educational data mining

Structured for faculty review before reading abstracts

Explainable AI Time-Series Forecasting Edge Inference Educational Data Mining

Top Indexed Venue

Q1 Journal

Artificial Intelligence Review under Springer Nature

Publication Mix

1 Journal + 1 IEEE Paper

2 additional manuscript-stage studies

Indexing Footprint

SCIE, Scopus, IEEE Xplore

Visible access path and venue class on every card

Professor Readout

Metrics First

Venue, ranking, and access appear before the paper details

IEEE
COMPAS 2025 IEEE Xplore indexed

IEEE 2nd International Conference on Computing, Applications and Systems

International conference proceedings paper with full-paper publication access.

Type Conference
Format Full Paper
Access IEEE Xplore

Bridging Climates: Weather Forecasting Using OLS and ARIMA Models with a Multi-Country Dataset

Authors:

Arafat Hossain, Md. Salauddin Sarker, Aditya Rajbongshi, Most. Rakiba Khanom Jisa, Maria Afrin Bindu, Kayes Mohammad Abdullah

Key Contribution

Conducted a 61-year multi-country clinical analysis of climate data, proving OLS outperforms ARIMA (RMSE 0.30) for baseline long-term predictions in Bangladesh.

Modeling Approach

OLS + ARIMA hybrid forecasting on a 61-year multi-country climate dataset (Bangladesh, Saudi Arabia, Japan, Russia).

Research Impact

OLS outperformed ARIMA for Bangladesh (RMSE 0.30, MAE 0.25); accurate long-term weather prediction via time-series analytics.

IDAA
IDAA 2025 Scientific Forum

Industrial and deployment-focused AI forum submission

Specialized track manuscript positioned for industrial edge inference discussion.

Type Forum Track
Status Manuscript
Access On Request

EdgeVision: Quantization-Aware Inference for Industrial Edge

Authors: Md. Salauddin Sarker, et al.

Key Contribution

Developed an 8-bit INT quantization approach for ARM edge nodes achieving sub-10ms inference latency targeting industrial vision.

Optimization Focus

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

Research Position

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

EDM
Educational Data Mining Accuracy 94.74%

Applied educational AI manuscript with explainability layer

Research-stage paper focused on retake prediction and interpretable student-risk modeling.

Type Research Paper
Status Manuscript
Access On Request

Bridging Educational Gaps: Predicting Retakes Among Bangladeshi Undergraduates with Machine Learning and Explainable AI

Authors:

Nusrat Zahan Nila, Md. Ali Mahmud Pritom, Fatema Tuz Johora, Aditya Rajbongshi, Md. Salauddin Sarker, Md. Ashrafuzzaman

Key Contribution

Built an early course-retake prediction framework for Bangladeshi undergraduates using a soft voting classifier with SHAP and LIME, reaching 94.74% accuracy on a 474-student dataset.

Predictive Framework

474 real student records, extensive preprocessing, and baseline comparison across LR, GNB, RF, SVC, Perceptron, KNN, and a proposed soft voting ensemble.

XAI Insights

SHAP and LIME exposed the factors driving retake risk, including lack of preparation, family issues, poor time management, and related academic pressures.

Verified Credentials

Technical Validation.

Academic Certifications & Professional Recognition

8 Certificates
3 Conferences
Dean's List
Verified
Status

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.

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

Bridging Climates: Multi-Country Weather Forecasting

OLS vs ARIMA evaluation across 61 years of multi-country climatic data (Bangladesh, Saudi Arabia, Japan, Russia).

The Problem

Forecasting weather variables accurately across geographically diverse regions.

The Data

61 years of historical weather data across multiple regions.

Approach: OLS + ARIMA hybridization; comprehensive error diagnostics and baseline comparisons across 4 countries.

Metrics: OLS outperformed ARIMA for Bangladesh (RMSE 0.30, MAE 0.25); demonstrated reliable long-term prediction via time-series analytics.

KEY INSIGHT (LEARNING)

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

Weather Forecasting
Multi-Country Climate Analysis
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.

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 Published Papers 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.

Get in Touch.

Collaboration & Inquiry

I am always open to discussing new opportunities in Explainable AI, Industrial Automation, and Odoo ERP. Whether you have a question or just want to say hi, I'll try my best to get back to you!

Academic Referees

Available upon request for MSc applications.

Location

Dhaka, Bangladesh

WhatsApp / Phone

+880 1521-260707

Message Prepared

Your email client has been opened with the message prefilled. Just hit send!

Submitting opens your email app with this message prefilled.