TaskManager - Enterprise Task Management Platform
22
TaskManager - Enterprise Task Management Platform


🚧 Currently in active development — Live demo coming soon!

Building a production-ready, full-stack task management system designed with enterprise-grade architecture from the ground up. While it may seem overengineered for a "simple" task manager, the goal is to demonstrate how to build scalable SaaS infrastructure that could support features like multi-tenant team workspaces, user-based rate limiting (similar to Microsoft Teams), and enterprise collaboration — all ready for growth.

💭 Motivation:
This project serves as a technical portfolio piece proving senior-level proficiency in modern web development, microservices architecture, and DevOps practices. Every architectural decision is intentional and production-ready, designed to scale from a personal tool to a full SaaS platform without rewriting the foundation.

Real-time Communication:

  • Laravel Reverb WebSocket server for instant task updates and live notifications
  • Presence channels showing who's online in real-time
  • Desktop and in-app notification system with configurable preferences

Authentication & Security:
  • Multi-factor authentication: Passkeys (WebAuthn biometric), TOTP 2FA, and traditional credentials
  • SSO integration with Google, Microsoft, GitHub, and Apple OAuth providers
  • Laravel Sanctum API tokens with rate limiting and brute-force protection
  • Team-based RBAC (Role-Based Access Control) for enterprise authorization scenarios

Task Management Features:
  • Full CRUD with categories, flexible tagging, file attachments, and collaborative comments
  • Recurring tasks (daily, weekly, monthly, yearly) with subtask breakdown
  • Due dates, reminders, and activity logs archived to MongoDB for analytics
  • Meilisearch-powered full-text search with typo tolerance, faceted filtering, and instant autocomplete

Hybrid API Architecture:
  • RESTful API with OpenAPI/Swagger interactive documentation
  • GraphQL API (Lighthouse) with GraphiQL playground for flexible queries
  • Dual API strategy: REST for CRUD simplicity, GraphQL for complex dashboard aggregations

Observability & Monitoring:
  • Sentry integration for error tracking and performance monitoring
  • Laravel Telescope for request debugging and Laravel Horizon for queue monitoring
  • Redis caching with configurable TTL for dashboard statistics
  • Health check endpoints and CSP headers with nonce support

Production Infrastructure:
  • 18 Kubernetes manifests with HPA auto-scaling, network policies, and ingress configuration
  • PostgreSQL (primary ACID DB), MongoDB (activity logs), Redis (cache/pub-sub), RabbitMQ (message queue)
  • 3 deployment profiles: Minimal (optimized for 4GB RAM demo VPS), Standard (production), Enterprise (K8s HA)
  • Graceful degradation: RabbitMQ falls back to Redis queues, Meilisearch falls back to DB search

DevOps & Quality:
  • 7 CI/CD GitHub Actions workflows with automated testing and security scanning (Trivy)
  • Pest PHP testing framework with 120+ test cases and code coverage via Codecov
  • k6 load testing suite (smoke, load, stress, spike scenarios up to 500 concurrent users)
  • Automated screenshot generation via Playwright on each release
  • PHPStan Level 8 static analysis, Pint/PHP-CS-Fixer code styling

Internationalization:
  • Full i18n support with 3 languages: English, German, Turkish
  • Vue I18n integration for frontend localization

Laravel
Vue
TypeScript
PHP
PostgreSQL
MongoDB
Redis
RabbitMQ
Meilisearch
Docker
Kubernetes
GraphQL
REST API
CrossPacket - Multi-Language Packet Generator
CrossPacket - Multi-Language Packet Generator

Built a cross-platform data packet generator that generates type-safe serialization code for 9 programming languages from a single JSON schema definition. Define packets once, generate everywhere — optimized for WebSocket communication, real-time applications, and microservices requiring consistent data structures across polyglot tech stacks.

Supported Languages & Libraries:

  • C# - System.Text.Json (built-in) + MessagePack-CSharp
  • C++ - yyjson + msgpack-c libraries
  • Dart/Flutter - dart:convert (built-in) + msgpack_dart
  • Go - encoding/json (built-in) + vmihailenco/msgpack
  • Java - Jackson + msgpack-java (Maven/Gradle compatible)
  • PHP - PSR-12 compliant with json (built-in) + msgpack-php
  • Python - PEP 8 compliant with json (built-in) + msgpack-python
  • Rust - serde_json + rmp-serde for serialization
  • TypeScript - JSON (built-in) + @msgpack/msgpack

Core Features:
  • Dual serialization: JSON for debugging/interoperability, MessagePack for performance
  • JSON Schema validation with packets.schema.json for packet definitions
  • Field-level validation: regex patterns, bounds checking, required fields, type enforcement
  • Security utilities for Python and C# with configurable size limits and input validation
  • Idiomatic code generation following each language's official style guidelines
  • Deep map conversion for safe handling of nested msgpack structures
  • Timezone-aware datetime with ISO 8601 format
  • Configurable type_field for packet type identification
  • Optional formats via --no-msgpack or --no-json CLI flags

Generated API (Consistent Across All Languages):
  • Default and parameterized constructors for flexible initialization
  • toJson() / toMsgPack() for serialization
  • fromJson() / fromMsgPack() static methods for deserialization
  • Getters/setters with language-appropriate naming conventions

Testing & Quality:
  • Comprehensive test suites for all 9 languages with 95%+ coverage target
  • GitHub Actions CI testing all languages in parallel with Codecov integration
  • Edge case coverage: integers (INT_MAX/MIN), floats (NaN, Infinity), nested structures, unicode
  • Canonical test data ensuring identical serialization behavior across all implementations

Ideal Use Cases:
  • WebSocket communication between mobile apps and backend servers
  • Microservices with polyglot architecture needing shared data contracts
  • Real-time gaming, IoT, and high-performance applications
  • Mobile apps with bandwidth constraints (MessagePack is more compact than JSON)

Python
Dart
TypeScript
Java
Rust
Go
C++
C#
PHP
MessagePack
JSON Schema
ReceiptSnapper - AI-Powered Receipt Management
26
ReceiptSnapper - AI-Powered Receipt Management


🚧 Currently in active development

Developing a comprehensive cross-platform personal finance and receipt management application using Flutter. The app combines AI-powered OCR technology for automatic receipt data extraction with full budget tracking, spending analytics, and German-compliant bookkeeping features (GoBD/DATEV export).

OCR & Image Processing:

  • Multi-engine OCR pipeline: Google ML Kit (mobile), Tesseract (desktop)
  • YOLOv8 receipt boundary detection via TensorFlow Lite with automatic cropping for improved OCR accuracy
  • Platform abstraction layer enabling unified API across Windows, macOS, Linux, iOS, and Android
  • Intelligent receipt parsing with pattern recognition for amounts, dates, vendors, and VAT rates

Finance & Budget Features:
  • Full budget management with income/expense tracking and recurring transactions
  • Interactive statistics with multiple chart types (pie, bar, line, area, stacked)
  • Smart spending insights with weekly/monthly summaries and anomaly detection
  • Tax report generation with German tax category support (SKR03/04 accounts)
  • DATEV-compatible export for professional accounting integration

Security & Data Management:
  • PIN and biometric authentication with hardware-backed encryption (Android Keystore/iOS Keychain)
  • PBKDF2 key derivation (100,000 iterations) with tamper detection and auto-wipe protection
  • Multi-format export: CSV, Excel, PDF, JSON, and SQLite database backup
  • GoBD-compliant audit trail logging for German bookkeeping requirements

Cloud Sync (Architecture Complete, Implementation Planned):
  • Designed multi-provider cloud sync architecture supporting Google Drive, OneDrive, Dropbox, and iCloud
  • Append-only sync strategy for safe multi-device and shared account usage
  • Automatic conflict resolution with merge, local-wins, and cloud-wins strategies

Technical Stack:
  • Drift (SQLite) with type-safe queries and automatic schema migrations
  • fl_chart for interactive financial visualizations
  • Full i18n support (German/English) with flutter_localizations
  • CI/CD with GitHub Actions for automated testing and release builds

Flutter
Dart
TensorFlow Lite
YOLOv8
Google ML Kit
Tesseract OCR
SQLite
Personal Website

Designed and developed a responsive portfolio website showcasing technical projects and full-stack capabilities.

Implemented an interactive 3D background using Three.js with custom GLSL shaders to render animated particle systems and dynamic visual effects during the launch sequence. Optimized rendering performance to maintain 60 FPS across devices while ensuring responsive design principles. Built with React Router for seamless navigation and Material-UI for a modern, accessible interface.

React
Three.js
JavaScript
HTML
CSS
Survey Platform with Load Balancing
5
Survey Platform with Load Balancing

Built a containerized survey application enabling instant QR code–based survey creation and anonymous participation. University project for Cloud Computing course.

Set up two Spring Boot instances behind HAProxy load balancer with health checks and round-robin distribution. Implemented real-time result visualization using Chart.js, admin dashboard for survey management, and PostgreSQL for persistent data storage.

Technical Implementation:

  • Containerized all services using Docker with network isolation between app and database layers
  • Multi-stage Docker builds producing optimized ~300MB images
  • GitHub Actions CI/CD: 87 tests with 80% coverage threshold (JaCoCo)
  • Semantic versioning with automated releases to GitHub Container Registry
  • Spring Boot Actuator health monitoring integrated with HAProxy health checks

Java
Spring Boot
PostgreSQL
Docker
Docker Compose
HAProxy
Maven
Thymeleaf
Chart.js
Skribbl.io Friend Finder

Developed a Chrome browser extension to enhance multiplayer game functionality by automating friend discovery.

Implemented DOM manipulation and real-time monitoring to detect and join game sessions containing specified users. Utilized Chrome Extension APIs including background scripts, content scripts, and message passing for seamless browser integration.

Technical Features:

  • Asynchronous session scanning with configurable search parameters
  • Client-side filtering and matching algorithms for efficient friend detection
  • User-friendly popup interface with persistent settings storage

Note: Functionality may be affected by recent updates to Chrome's Manifest V3 requirements and changes to Skribbl.io's architecture.

Chrome Extension
JavaScript
Real-Time Object Detection in Red Dead Redemption 2
Real-Time Object Detection in Red Dead Redemption 2

Developed a computer vision system for real-time object detection in video game environments using YOLOv8.

Manually annotated and curated a custom dataset using Roboflow, then trained a specialized model to recognize in-game entities with high accuracy. Implemented multi-threaded processing pipeline to maximize detection throughput and minimize latency during live gameplay analysis.

Technical Approach:

  • Dataset creation and labeling workflow using Roboflow's annotation platform
  • Model training on both AMD (ROCm) and NVIDIA (CUDA) architectures for performance comparison
  • Real-time visualization framework using matplotlib for detection overlay and confidence metrics
  • Thread-safe frame capture and inference pipeline with configurable batch processing

Performance is hardware-dependent, with GPU acceleration significantly improving inference speed and frame throughput.

Python
YOLOv8
OpenCV
Multi-threading
matplotlib
supervision
ROCm
CUDA
FlapPyBlink - Eye-Controlled Game

Developed an innovative human-computer interaction system integrating computer vision with game mechanics.

Implemented OpenCV-based eye-blink detection algorithms to replace traditional input controls in a Python implementation of Flappy Bird. Collaborated with a peer to design and optimize the detection pipeline for real-time responsiveness and accuracy.

Technical Implementation:

  • Real-time facial landmark detection using Haar Cascades and eye aspect ratio (EAR) calculations
  • Calibration system for personalized blink detection thresholds
  • Low-latency input processing pipeline to ensure responsive gameplay
  • Frame buffering and smoothing algorithms to reduce false positives

Python
OpenCV
CodeAcademy - Interactive Learning Platform
9
CodeAcademy - Interactive Learning Platform

Collaborated with a three-person development team to create an interactive coding education platform emphasizing hands-on learning over traditional theory-heavy approaches. Developed a full-stack application using Laravel-React-Mix with MySQL, contributing to system design, core implementation, and team coordination.

Platform Features:

  • Isolated code execution environment using Docker containers for secure multi-language support (Java, JavaScript, Python)
  • Dynamic content management system with HTML parsing for rich lesson formatting, syntax highlighting, and embedded media
  • Automated assessment engine comparing student output against expected results with configurable tolerance
  • Gamification system tracking user progress through XP, achievements, and visual statistics dashboards

Technical Architecture:
  • RESTful API communication layer using Axios for client-server interaction
  • Token-based authentication and session management
  • Container orchestration for on-demand code execution with resource limits and timeout controls
  • Responsive frontend with syntax-highlighted code editor and real-time output display

Laravel
React
Docker
PHP
MySQL
Java
JavaScript
Python
University Appointment Scheduling System
8
University Appointment Scheduling System

Developed a university-wide appointment scheduling platform enabling students and staff to book consultation time slots with lecturers. Built a Laravel-React application with MySQL backend as part of a three-person team responsible for requirements analysis, system design, and implementation.

Key Features:

  • Dual appointment system supporting both recurring availability slots and custom one-time appointments
  • Email-based authentication for students using university domain verification with rotating token-based access
  • Hierarchical admin system allowing lecturer promotion and delegation of administrative privileges
  • Integrated email notifications with configurable preferences for appointment confirmations and reminders

Technical Implementation:
  • Interactive calendar interface built with DevExpress React Scheduler with configurable business hours
  • RESTful API architecture using Axios for client-server communication
  • Server-side time zone handling and conflict detection using Laravel's Carbon library
  • Secure hash-based token system for passwordless student authentication

Laravel
React
PHP
MySQL
JavaScript
HTML
CSS
RESTful API
MCBedrockQuery - Minecraft Server Query Tool
3
MCBedrockQuery - Minecraft Server Query Tool

Developed a cross-platform solution for querying Minecraft: Bedrock Edition servers using the game's proprietary UDP-based query protocol. Implemented low-level network communication to retrieve server information including MOTD (Message of the Day), player count, version details, and latency metrics.

Technical Implementation:

  • Custom UDP packet construction and parsing for Bedrock Edition's query protocol
  • Java implementation providing cross-platform compatibility and CLI interface
  • Android port leveraging Java codebase for native mobile query execution
  • Asynchronous network operations with timeout handling and connection retry logic

Java
Android