Projects
A collection of projects I've built and contributed to.
MAIN PROJECTS
Stephen's Privélessen Platform
Digital Tutoring Platform (In Development)
- Digital platform for private tutoring business built with Next.js 14 and TypeScript
- Currently used to distribute notes and materials to students and parents with responsive design
- Modern web architecture with component-based frontend and API integration capabilities
2023 - Present
Personal Website
Next.js 14 Application
- Production website built with Next.js 14 and TypeScript, deployed live at stephenadei.nl
- Bilingual content (EN/NL) with AI-driven note-taking features and modern UX principles
- Responsive design with SEO optimization and performance best practices
2023 - Present
WhatsApp Bot
Automated Scheduling System
- Automated scheduling and FAQ system prototype built with Node.js and WhatsApp Business API
- Integration with existing Next.js tutoring platform for seamless booking workflow
- Natural language processing for intelligent conversation handling
2024 - Present
StephenStat
Web-Based Statistical Analysis Platform
- Web-based statistical analysis platform built with Next.js, TypeScript, and modern web technologies
- Structured guidance system with automatic assistance at each analysis step and variable recognition
- Automatic data upload, measurement level configuration, and intelligent analysis recommendations
2024 - Present
digibinas
BINAS Digital API & Data Platform
- Full-stack application with modular Node.js and Express backend API serving Dutch BINAS educational data
- Modern JSON-based data structure with ETL pipeline for data optimization and transformation
- Multi-repo monorepo architecture with automated deployment scripts and Docker containerization
2023 - Present
Amsterdam Street Furniture Explorer
Enterprise Frappe Framework Application
- Enterprise Frappe Framework v15 application that imports and synchronizes street furniture data from Amsterdam open data APIs
- Automated daily synchronization of 5+ different data sources with Python-based data integration logic
- Interactive Leaflet.js-based map with clustering, filtering, and geospatial visualization features
2023 - Present
Ohpen Data Engineering Business Case
Production-Grade Data Platform Architecture
- Data engineering case study (full documentation at ohpen.stephenadei.nl): finance-grade data platform architecture for fintech/banking, designed and documented end-to-end
- Bronze/Silver/Gold medallion architecture (AWS S3) with schema versioning and governance workflows for analytical (OLAP) workloads
- Production ETL pipeline (Python/PySpark) with comprehensive validation, error handling, quarantine management, and condemned data lifecycle
2025 - Present ohpen.stephenadei.nl
Analog Photography Portfolio
Production Photography Portfolio Website
- Live production photography portfolio website built with Next.js 13+ and TypeScript, optimized for performance and SEO
- Responsive masonry layout design with Framer Motion animations and Tailwind CSS for modern UI/UX
- Image optimization and CDN integration using Cloudinary for fast loading and bandwidth efficiency
2023 - Present
Ableton AI Tools
Cross-platform Desktop Application
- Started with open source ableton-mcp project by Siddharth Ahuja, providing Model Context Protocol integration for Ableton Live
- Comprehensive collection of AI-powered tools: Python MCP server with 28+ tools, Electron/React desktop GUI, and standalone .als file analyzer
- Enables natural language control of Ableton Live through AI assistants (Claude Desktop, Cursor, or custom GUI)
2024 - Present
TUTORING & CONSULTING PROJECTS
Academic guidance projects demonstrating technical depth across multiple domains.
Operations Research / Simulation (Java)
Java discrete-event simulation for service systems
- Guided discrete-event simulation project in Java for service systems (hospital/ambulance), including event logic and object-oriented modeling of entities (accident/ambulance/hospital/region)
- Supported implementation and analysis of queueing/threshold-queue simulations to compare system behavior under different policy parameters (performance/throughput as function of thresholds)
- Modeled the newsvendor inventory problem and used simulation/expectation analysis to evaluate decision rules (order quantity) on profit/risk trade-offs
discrete-event simulationqueueing theoryinventory optimizationMonte CarloJava OOP
Security & Machine Learning
Deep learning for security use-cases (SS7 attack detection, phishing detection)
- Supported research and development of deep learning classification models for security use-cases (including SS7 attack detection) with focus on practical detection in noisy/real-world data
- Guided design of phishing URL detection pipeline: dataset ingestion (CSV), preprocessing/feature construction, training & evaluation of models for binary classification
- Experimented with semi-supervised learning (labeled + unlabeled) to improve performance when labels are scarce
- Documented features and model inputs for reproducibility and transferable interpretation toward stakeholders
deep learningsemi-supervisedclassificationfeature engineeringcybersecurity analytics
Data Processing & Representations (Python)
Python ETL pipeline, experiment automation, reproducible workflows
- Built a multi-stage data processing pipeline in Python (phase-based scripts) including reading/cleaning/transformations of CSV data and structured exports
- Automated experiment runs (parameter sweeps) and captured results in reusable tables/files for comparison and analysis
- Created visualizations (Jupyter + plots) to make trends and outcomes per phase/parameter insightful and justify choices
- Wrote documentation per phase to make the workflow reproducible and transferable
PythonPandas/NumPyETLreproducible pipelinesexperimentationJupyter
LISS panel data (mixed-effects / longitudinal)
Longitudinal panel data analysis with mixed-effects models
- Analyzed longitudinal panel data (multi-wave LISS; CSV/SAV) with attention to consistent merges, variable definitions (codebook) and data quality
- Built and interpreted mixed-effects (logistic) regression models to estimate effects in hierarchical/longitudinal data
- Developed an end-to-end analysis workflow (data cleaning → model fitting → reporting) with repeatable steps for transparent decision-making
panel datamixed-effectslogistic regressioncausal-ish modeling disciplinereproducibility