Loading…
Building data systems,models that .
Undergraduate Software Engineering student at Daffodil International University and Junior Data and Research Associate Intern at Research and Management Consultants Ltd. (RMCL), working on time-series forecasting, data engineering, public health analytics, and full-stack systems through reproducible applied projects.
Meherab Hossain Shafin
PythonSQLNext.jsForecastingPublic Health Analytics
0+Deployed Web PlatformsRMCL Procurement Portal, RMCL website, LTC website
0M+Records Processed via ETLNYC taxi analytics pipeline
0.00sOptimized Query LatencyReduced from about 1.2s
0+Applied Forecasting ProjectsDengue, energy, air quality, and more
CapabilitiesApplied
Applied
capabilities.
Concrete systems built through forecasting, data engineering, full-stack delivery, and validated analytical workflows.
01
Forecasting & Statistical Modeling
Built forecasting workflows using SARIMA, Random Forest, Gradient Boosting, DeepAR, and benchmark models under rolling-origin and walk-forward validation.
0+applied forecasting projects
Workflow
Explore.Validate.Deploy.
From raw data to reliable models and usable systems.
Stack
Engineering systems built for reproducibility.
From data ingestion and temporal validation to mobile applications, institutional platforms, and production deployment.
Data Engineering
ETL pipelines, schema design, migrations, indexing, API ingestion, and analytical databases.
PythonSQLPostgreSQLMySQLSQLiteasyncpgAlembicPrismaAPI ingestionSchema validationQuery optimization
Forecasting and Validation
Time-series modeling, baseline comparison, leakage prevention, and rolling-origin evaluation.
SARIMARandom ForestGradient BoostingDeepARRegressionWalk-forward validationRolling-origin evaluationBaseline comparisonForecast-horizon evaluation
Application Engineering
Mobile, web, API, and institutional systems with authentication and real-time communication.
Next.jsReactReact NativeExpoFastAPILaravelTypeScriptPHPREST APIsWebSocketsPydanticZodZustand
Reliability and Deployment
Testing, CI, migrations, environment validation, HTTPS, monitoring, and release workflows.
pytestPlaywrightGitHub ActionsLinuxHTTPSEnvironment validationDatabase migrationsHealth checksSentryDeployment documentation
MethodologyValidated workflows,
Validated workflows,
not vanity metrics.
Quantitative proof points live in the hero and flagship project cards. This section summarizes how work is evaluated and delivered.
- Walk-forward validation
- Rolling-origin evaluation
- Reproducible pipelines
- Decision-support systems
PrinciplesEngineering
Engineering
principles.
Reliable systems require validation, reproducibility, performance awareness, and interpretable results—interconnected components of one unified engineering philosophy.
Leakage Prevention
Temporal validation and defensive modelling
Reproducibility
Version-controlled, repeatable workflows
Performance Awareness
Efficient queries, caching, and production readiness
Interpretability
Explainable outputs and transparent comparisons
RELIABLEBY DESIGN
Leakage Prevention
Temporal validation and defensive modelling
Reproducibility
Version-controlled, repeatable workflows
Performance Awareness
Efficient queries, caching, and production readiness
Interpretability
Explainable outputs and transparent comparisons
ProjectsSelected
Selected
projects.
Applied projects with documented outputs across forecasting, data engineering, environmental analytics, and deployed web systems.
01Flagship Data Engineering Project
NYC Taxi Data Engineering
Engineered a PostgreSQL ETL and analytics pipeline for more than 40 million NYC yellow taxi records.
0M+records
0.00squery latency
Key output: Reduced time-filtered query latency from about 1.2 seconds to 0.05 seconds through schema design and indexing.
Data EngineeringPostgreSQLSQL Analytics
02Flagship Public Health AI Case Study
DengueOps AI
An experimental dengue forecasting and dataset-assessment platform designed to validate temporal data, support dataset-specific model evaluation, and translate analytical outputs into planning-oriented evidence.
Key output: Separated rapid forecasting from deeper dataset assessment through Quick Forecast and Assess Dataset workflows, while defining canonical planning outputs and explicit model limitations.
Public Health AIForecastingTime-SeriesModel Assessment
03Flagship Forecasting Project
Electricity Load Forecasting for Ireland
Developed a multi-horizon electricity demand forecasting pipeline using ENTSO-E hourly demand and NASA POWER weather data.
Key output: Benchmarked seasonal naive, SARIMAX, quantile gradient boosting, weather-augmented gradient boosting, and DeepAR across t+1, t+24, and t+168 horizons.
EnergyForecastingMachine Learning
04
Bangladesh Dengue Forecasting
Built a reproducible 15-year national dengue forecasting workflow using surveillance data and lagged climate variables.
Key output: Compared naive persistence, autoregressive regression, SARIMA, and Random Forest under strict walk-forward validation.
ForecastingPublic HealthTime-Series
05Flagship Full-Stack Booking Platform
DIU Turf
A mobile-first turf booking platform for hostel students of Daffodil International University, combining fair-use reservations, waitlists, real-time updates, and role-based administration.
0Active users
0Backend tests
Key output: Implemented transaction-safe booking, automatic waitlist promotion, authenticated real-time updates, and an analytics-ready PostgreSQL data model for a live system with 26 active users.
React NativeFastAPIPostgreSQLBooking System
06Flagship Institutional Workflow Platform
DIU Internal Workflow Management System
A role-based institutional platform for Daffodil International University, coordinating projects, work items, tasks, communication, progress reporting, repositories, approvals, and accountability across administrative levels.
0System use cases
0Primary system roles
Key output: Designed a project-linked operational workflow that separates managerial coordination from subordinate execution while preserving reporting, communication, attachment history, role boundaries, and institutional accountability.
LaravelMySQLRBACWorkflow Management
07Flagship Full-Stack Platform
RMCL Official Platform
Developed a secure full-stack organizational web platform using Next.js, PostgreSQL, Prisma ORM, CMS workflows, authentication, and caching.
0scache revalidation
Key output: Implemented API validation, session authentication, asset upload safeguards, ISR/database caching, and deployment documentation.
Next.jsFull-StackPostgreSQL
08
Dataset Reliability Inspector
Built a deterministic dataset auditing system for schema integrity, missingness, duplicates, outliers, leakage signals, and modeling readiness.
Key output: Implemented transparent 0–100 readiness scoring using rule-based penalties instead of black-box scoring.
Data QualityValidationModeling Readiness
09
Bangladesh Air Quality Forecasting
Built a PM2.5 forecasting workflow across eight Bangladeshi cities using automated API ingestion and SQLite storage.
0cities modeled
Key output: Compared ARIMA and SARIMA models with rolling-origin validation to identify city-level forecasting behavior.
Environmental AnalyticsForecastingPublic Health
01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001 01010101 00110011 11001100 10101010 01100110 10011001
Research
Research focus.
My research direction focuses on validated forecasting workflows, public health analytics, environmental data, and reproducible computational methods. Current outputs include preprint-style manuscripts and applied modeling pipelines.
Research interests
Time-Series ForecastingPublic Health ModelingEpidemiological AnalyticsEnvironmental AnalyticsData EngineeringReproducible Computational Methods
Outputs
- Status: Preprint manuscript
- Status: Preprint manuscript
Co-curricular
Co-curricular highlights.
Selected debating, ambassadorial, and academic recognition activities beyond technical project work.
Inter Cantonment National Debate Competition
Runner-up, 2024
Campus Ambassador
Green Genesis Fest, 2026
DIU Division of Research Poster Presentation
Certificate of Achievement for Electricity Load Forecasting Model
AUST DC National Debate Competition
Runner-up, 2023
Adamjee School Environmental Debate
Champion, 2018
Contact
Let's build analytical systems that matter.
Open to MSc research opportunities, data science collaborations, forecasting projects, internships, and full-stack development work.
- Email:meherabhossainshafin@gmail.com
- GitHub:github.com/MeherabHS
- LinkedIn:linkedin.com/in/meherab-hossain-3b3416403
- Facebook:facebook.com/mrcube.16
- Location:Dhaka, Bangladesh
Send an inquiry
Collaborations, research, internships, or project opportunities. Messages are delivered to meherabhossainshafin@gmail.com.