WELCOME TO MY PORTFOLIO

Hello, I'm

Md. Masum
Billah

Masum
Open to Opportunities

About Me

I’m Md. Masum Billah, an Executive Engineer specializing in AI & Mobile Apps and the Founder of SpideyTech. As a Computer Science & Engineering graduate, backed by a strong foundation in data structures, algorithms, and competitive programming, I engineer production-grade mobile applications, full-stack web platforms, AI automation workflows, and applied machine learning systems.

My core development stack spans Flutter, React Native, React.js, Next.js, Django REST Framework, FastAPI, and Laravel, supported by real-time, containerized architecture and automation using Docker, PostgreSQL, Redis, Django Channels, and n8n. Through platforms such as CricketOps, MessiFy, Spinora Music, and SiliconSense, I connect sleek user interfaces with low-latency backend pipelines, automated workflows, and scalable databases.

In the research space, I have authored two Q2-published research papers on hyper-local air pollution modeling and traffic sign recognition. Combining ensemble learning, computer vision, custom datasets, and Google Earth Engine (GEE), I turn applied research into production-ready software built for real-world impact.

Technical Skills

Mobile & Web

Flutter React Native React.js Next.js Dart JavaScript TypeScript Tailwind CSS Riverpod

Backend & Architecture

Django REST Framework FastAPI Laravel Django Channels / WebSockets REST API Architecture

AI, Vision & Research

Computer Vision (YOLO, ResNet, RT-DETR) Deep Learning & Ensembles XAI (SHAP, LIME) Google Earth Engine (GEE) Scikit-Learn PyTorch OpenCV Pandas & NumPy

DevOps, DB & Automation

Docker PostgreSQL Redis n8n Automation Firebase MySQL SQLite / Drift Linux / Ubuntu Git & GitHub

Core CS & Hardware

C++ Python Data Structures & Algorithms Competitive Programming Codeforces LeetCode IoT & Networking

Projects

HybridVision: Real-Time Traffic Sign Detection & Classification

HybridVision: Real-Time Traffic Sign Detection & Classification [YOLOv8 + ResNet50]

Custom Bangladeshi traffic sign dataset (9,133 images, 29 classes) + Streamlit web app for real-time detection & classification using a hybrid pipeline: YOLOv8 (detection) + ResNet50 (classification).

Computer Vision Deep Learning Custom Dataset Creation Annotation (Roboflow) Bangladeshi Road Signs 9133 Images 29 Classes Hybrid Model Traffic Sign Detection Traffic Sign Classification YOLOv5 YOLOv7 YOLOv8 YOLOv11 ResNet50 Model Comparison mAP@0.5-100% Precision-100% Recall-99% Accuracy-99% Roboflow PyTorch OpenCV Streamlit Web App Image Upload Video Upload Live Camera Kaggle GPU(P100) Real-Time Inference Deployment Ready Pickle / PKL ONNX TensorRT TFLite CARLA Simulation Autonomous Driving
CricketOps: Real-Time Cricket Scoring, Tournament Management & Broadcast Platform

CricketOps: Real-Time Cricket Scoring, Tournament Management & Broadcast Platform

Full-stack cricket operations platform featuring ledger-based ball-by-ball scoring, tournament management, Flutter mobile scoring, offline synchronization, WebSocket live updates, and OBS broadcast overlays powered by Django, PostgreSQL, Redis, and Flutter.

Full Stack Development Cricket Scoring System Tournament Management Real-Time Application Ball-by-Ball Scoring Ledger-Based Architecture Event-Driven Architecture Scoring Engine Django Django REST Framework Django Channels Django ASGI REST API WebSocket PostgreSQL Redis Flutter Mobile App Offline-First Offline Sync Idempotency Database Transactions Row-Level Locking Optimistic Versioning Audit Trail Score Correction Live Score Live Match Updates Match Statistics Player Statistics Batting Statistics Bowling Statistics Partnership Tracking Fixtures Points Table Net Run Rate T20 ODI Limited Overs Cricket Configurable Cricket Rules OBS Integration PRISM Live Studio Broadcast Overlay Live Streaming Docker Docker Compose Scalable Architecture Single Source of Truth Production-Oriented Architecture
Spinora Music: Flutter Android music player with offline and online playback

Spinora Music: Offline & Online Android Music Player [Flutter + Dart]

Flutter Android music player with local and online playback, synced lyrics, equalizer, crossfade, Smart Mixes, playlists, and dark/light glassmorphism themes.

Flutter Dart Android Development Music Player Offline Playback Online Music Search Background Playback Notification Controls Queue Management Playlists & Favorites Synced Lyrics LRC Parser LRCLIB API Audio Equalizer Loudness Enhancement Crossfade Sleep Timer Playback Speed Smart Mixes Glassmorphism UI Dark & Light Themes Riverpod go_router Drift / SQLite just_audio audio_service Android MediaStore Kotlin Platform Channels Metadata Editing Backup & Restore Home-Screen Widget
Stroke Prediction System Using Machine Learning

Stroke Prediction System

Machine learning–based healthcare application for predicting stroke risk using patient medical and lifestyle data.

Machine Learning Healthcare Analytics Stroke Prediction Supervised Learning Binary Classification Random Forest Logistic Regression SMOTE Data Preprocessing Feature Engineering Model Evaluation SHAP Explainability Scikit-learn Python Joblib Model Deployment Healthcare ML End-to-End ML Project
MessiFy hostel and mess management app for meal tracking, expenses, deposits, and monthly settlement

MessiFy: Hostel & Mess Management System [Flutter + Django]

Full-stack hostel and mess management app with daily meal tracking, member deposits, expense ledgers, automatic balance calculations, and manager-led monthly settlement.

Flutter Dart Full Stack Development Mobile Application Hostel Management Mess Management Daily Meal Tracking Fractional Meal Counts Meal Rate Calculation Expense Ledger Member Deposits Shared Expense Splitting Balance Sheets Due & Surplus Tracking Monthly Settlement Cycle Locking Invite-Code Joining Manager & Member Roles Riverpod Dio Material 3 Light & Dark Themes Python Django Django REST Framework REST API JWT Authentication Token Refresh Flutter Secure Storage PostgreSQL Docker Compose n8n Integration Setup
SiliconSense by SpideyTech: PC performance, balance, bottleneck analysis, and upgrade guidance

SiliconSense by SpideyTech: PC Strength, Balance & Bottleneck Analyzer

Full-stack PC analyzer with workload-specific benchmark scoring, separate performance and balance insights, bottleneck detection, build comparison, and explainable upgrade guidance.

Full Stack Development PC Hardware Analysis Bottleneck Detection Workload-Based Scoring Performance & Balance Build Comparison Upgrade Recommendations Confidence Scoring Percentile Normalization Data Provenance Versioned Datasets Hardware Catalog Next.js React TypeScript Tailwind CSS Motion Recharts Python FastAPI Pydantic SQLAlchemy Alembic PostgreSQL Supabase Auth REST API Benchmark Ingestion Blender Open Data Phoronix Adapter Docker Pytest Vitest Glassmorphism UI Responsive Design
FoodHaven App

FoodHaven

Flutter-based food ordering & delivery app with Firebase integration.

Flutter Dart Firebase Mobile Application Food Delivery App OTP Authentication Phone Login Onboarding Screens Bottom Navigation Search Functionality User Profile Reusable Widgets Clean Architecture UI/UX Design Cross-Platform Firebase Core Responsive Design
Laptop Price Predictor

Laptop Price Predictor

Machine learning web app that predicts laptop prices based on specifications.

Python Machine Learning Regression Data Science Streamlit scikit-learn Pandas NumPy Model Deployment Predictive Analytics Feature Engineering Real-Time Prediction User Input Dashboard Pickle Model End-to-End ML Project Web Application
Smart IoT-Based Healthcare Monitoring Band

Smart IoT-Based Healthcare Monitoring Band

Wearable IoT healthcare device with real-time health monitoring and GPS tracking.

IoT ESP8266 Wearable Technology Healthcare Monitoring Heart Rate Monitoring SpO₂ Monitoring ECG Sensor Body Temperature GPS Tracking Real-Time Monitoring Blynk IoT Mobile App Wi-Fi Communication Embedded Systems Cloud Integration Remote Patient Monitoring Smart Medical Device IoT Healthcare
Online Banking Management System

Online Banking Management System

Java Swing–based banking system with MySQL database integration.

Java Java Swing MySQL JDBC OOP Authentication User Login System Transaction Management Balance Enquiry Deposit & Withdrawal PIN Management Database Design CRUD Operations Desktop Application Banking System
TrendyTreasure Online Gift Shop

TrendyTreasure – Online Gift Shop

PHP-based e-commerce web application with full admin and user functionality.

PHP MySQL JavaScript E-Commerce Product Catalog Shopping Cart Checkout System User Authentication Admin Dashboard Product Management Order Management CRUD Operations Session Handling Responsive Design Web Application
Six Seasons OpenGL Project

Six Seasons – OpenGL Seasonal Animation

C++ OpenGL project showcasing animated seasonal transitions using GLUT.

C++ OpenGL GLUT Computer Graphics 2D Animation Seasonal Simulation Timer-Based Animation Transformations Real-Time Rendering Interactive Graphics Keyboard Interaction Mouse Interaction Sun Animation Tree Growth Flower Blooming Snow Falling Effect Boat Movement Educational Project Graphics Programming
E-Learning Platform

E-Learning Platform

PHP & MySQL–based e-learning web application with multi-user roles.

PHP MySQL HTML CSS JavaScript Full Stack Role-Based Access User Authentication Student Dashboard Tutor Dashboard Admin Panel Video Streaming Course Playlists Like & Comment System Search & Filter CRUD Operations Responsive Design Web Application
Inventory and Sales Management System

Inventory & Sales Management System

Console-based C++ system for managing inventory and sales.

C++ Inventory Management Sales Processing File Handling Binary Files OOP Concepts Structs & Classes Menu Driven Console Application Stock Tracking Data Persistence Windows Console
Expense Tracker Application

Expense Tracker Application

Console-based C++ application for budget and expense management.

C++ Expense Tracking Budget Management File Handling Time-Based Logging Category Wise Expenses Console Application Menu Driven
Console Based Quiz Game

Console-Based Quiz Game

Menu-driven C++ quiz application with file handling.

C++ User Registration File Handling Multiple Quiz Segments MCQ Score Evaluation Replay Functionality

Research & Publications

My research explores how machine learning and computer vision can address practical challenges in air quality monitoring and road-scene understanding. These studies connect satellite observations, custom datasets, and hybrid modeling approaches with applications that make research accessible beyond the experimental environment.

Environmental ML 01

A Stacked Ensemble Learning Approach for Hyper-Local PM2.5 Prediction Using Satellite AOD and Meteorological Data

Md. Masum Billah et al. · First Author

Accurate local air quality estimation is challenging in areas with limited ground-based monitoring stations. This study investigates a stacked ensemble approach for predicting PM2.5 concentrations by combining satellite-derived Aerosol Optical Depth (AOD), meteorological variables, and geographic information, with Google Earth Engine (GEE) supporting the geospatial workflow.

Random Forest and XGBoost serve as complementary base learners, while a linear regression meta-model combines their predictions. The study evaluates the ensemble against individual models and examines which environmental features influence its estimates. A Flask-based API connects the trained model with live weather inputs to demonstrate PM2.5 prediction and AQI categorization, linking environmental research with an interactive application.

Google Earth Engine Satellite AOD Stacked Ensemble Random Forest XGBoost Linear Regression PM2.5 Prediction Flask API

Reported Results

Stacked Ensemble R²: 0.5592 Random Forest R²: 0.5367 XGBoost R²: 0.4276
Explore Research
Computer Vision 02

HybridVision: Real-Time Traffic Sign Detection & Classification

Md. Masum Billah et al. · First Author

Recognizing traffic signs on Bangladeshi roads requires handling varied backgrounds, small sign regions, and visually similar categories. HybridVision explores a two-stage deep learning pipeline: YOLO locates traffic signs, and ResNet50 classifies the cropped sign regions. The research is supported by a custom dataset of 9,133 images across 29 traffic sign categories, organized and annotated using Roboflow.

The workflow covers dataset preparation, augmentation, model experimentation, evaluation, and inference. A confidence-based fusion strategy combines detection and classification outputs into a final prediction. A Streamlit interface supports images, uploaded videos, and live-camera input, allowing users to explore the system beyond static experimental results. The work connects locally relevant dataset development with an accessible traffic sign recognition application.

YOLO ResNet50 Object Detection Image Classification Custom Dataset Roboflow Real-Time Inference Streamlit

Reported Results

YOLO mAP: 100% Detection Precision: 100% Detection Recall: 100% ResNet50 Accuracy: ~99%
Explore Research

Get In Touch

Have a project in mind or just want to chat? Feel free to reach out to me!