Steave Shabu
Computer science student with projects and practical experience in data analytics, machine learning, and AI projects. Professional experience in project management, IT support, and system administration.
Experience
• Salzgitter Mannesmann Handel: IT Working Student
• Milestone: Power BI Data Dashboard: Developed a data analytics dashboard by extracting and integrating data from Microsoft Defender, Intune, and Azure; applied ETL processes (cleaning, transformation, aggregation) to merge three data sources into a consolidated reporting layer. • Milestone: User Heatmap & Insights: Designed a global user heatmap and a user insights dashboard, enabling the identification of key insights and trends. • Tasks: Documentation, system configuration, support desk.
-
Power BI
Excel
Intune
Azure
• Aldi DX : IT Project Management Working Student
• Milestone: Predictive Intelligence: Developed, optimized, and tested a classification model for automated ticket assignment based on text data; responsible for model evaluation, training data quality, and performance optimization. • Tasks: Process and system documentation, platform optimization, cross-functional collaboration, international project work.
-
Predictive Intelligence
Machine Learning
Documentation
Project Management
• Scayan IT Working Student
• Milestone: Passbolt with Docker: Deployed a self-hosted Passbolt password management instance using Docker on a Linux server; improved credential security workflows for a team of fewer than 20 users. • Tasks: Software and hardware setup, IT support.
-
Linux
Docker
NAS Configuration
IT Support
Skills
Data Analytics and Machine Learning Projects
Federated Facial Paralysis Detection
Development of a privacy-preserving system for detecting facial paralysis based on federated learning. 3D facial landmarks were extracted using MediaPipe and their dimensionality was reduced via PCA before SVM and MLP classifiers were trained on distributed clients. Despite the medical data being distributed in a highly non-IID manner, the system achieved a global accuracy of approximately 60% and a local accuracy of up to 100% on individual clients.
- Federated Learning
- MediaPipe
- PCA
- SVM
- MLP
- Python
Berlin Airbnb Price Analysis
This project analyzes a dataset containing details of Airbnb bookings in Berlin. The goal is to understand how location, room type, and demand indicators influence pricing.
Car Price Analysis in Germany
This project focuses on insights and trends in a used car dataset based in Germany. A machine learning model that predicts car prices based on vehicle attributes has been built to provide accurate pricing predictions.
Chocolate Sale Analysis
This project analyzes sales performance for a chocolate company using Python, Pandas, and data visualization techniques. The goal is to uncover meaningful business insights by exploring revenue patterns, product performance, shipping behavior, and customer demand across countries and time periods.
Other Projects
Free Drive
A RC sized car that can be controlled with hand gestures. The hand gestures were trained and tested in Edge Impulse and then implemented with Arduino and ESP 32 on the car.
- GitHub
- Arduino
- Edge Impulse
- Team Management
- Leadership
- Team Play
DULAWS
A high rack maintenance software developed from scratch with a team of 22 students classified into frontend and backend developers. The project involved collaborative development, system design, and implementation of a comprehensive maintenance management solution.
- Team Play
- Collaboration
- Leadership
- Full Stack Development
- System Design